Mathias Unberath

Hao Ding, Jintan Zhang, Peter Kazanzides et al.

Vision-based segmentation of the robotic tool during robot-assisted surgery enables downstream applications, such as augmented reality feedback, while allowing for inaccuracies in robot kinematics. With the introduction of deep learning, many methods were presented to solve instrument segmentation directly and solely from images. While these approaches made remarkable progress on benchmark datasets, fundamental challenges pertaining to their robustness remain. We present CaRTS, a causality-driven robot tool segmentation algorithm, that is designed based on a complementary causal model of the robot tool segmentation task. Rather than directly inferring segmentation masks from observed images, CaRTS iteratively aligns tool models with image observations by updating the initially incorrect robot kinematic parameters through forward kinematics and differentiable rendering to optimize image feature similarity end-to-end. We benchmark CaRTS with competing techniques on both synthetic as well as real data from the dVRK, generated in precisely controlled scenarios to allow for counterfactual synthesis. On training-domain test data, CaRTS achieves a Dice score of 93.4 that is preserved well (Dice score of 91.8) when tested on counterfactually altered test data, exhibiting low brightness, smoke, blood, and altered background patterns. This compares favorably to Dice scores of 95.0 and 86.7, respectively, of the SOTA image-based method. Future work will involve accelerating CaRTS to achieve video framerate and estimating the impact occlusion has in practice. Despite these limitations, our results are promising: In addition to achieving high segmentation accuracy, CaRTS provides estimates of the true robot kinematics, which may benefit applications such as force estimation. Code is available at: https://github.com/hding2455/CaRTS

Zhenqi He, Mathias Unberath, Jing Ke et al.

Nuclei appear small in size, yet, in real clinical practice, the global spatial information and correlation of the color or brightness contrast between nuclei and background, have been considered a crucial component for accurate nuclei segmentation. However, the field of automatic nuclei segmentation is dominated by Convolutional Neural Networks (CNNs), meanwhile, the potential of the recently prevalent Transformers has not been fully explored, which is powerful in capturing local-global correlations. To this end, we make the first attempt at a pure Transformer framework for nuclei segmentation, called TransNuSeg. Different from prior work, we decouple the challenging nuclei segmentation task into an intrinsic multi-task learning task, where a tri-decoder structure is employed for nuclei instance, nuclei edge, and clustered edge segmentation respectively. To eliminate the divergent predictions from different branches in previous work, a novel self distillation loss is introduced to explicitly impose consistency regulation between branches. Moreover, to formulate the high correlation between branches and also reduce the number of parameters, an efficient attention sharing scheme is proposed by partially sharing the self-attention heads amongst the tri-decoders. Finally, a token MLP bottleneck replaces the over-parameterized Transformer bottleneck for a further reduction in model complexity. Experiments on two datasets of different modalities, including MoNuSeg have shown that our methods can outperform state-of-the-art counterparts such as CA2.5-Net by 2-3% Dice with 30% fewer parameters. In conclusion, TransNuSeg confirms the strength of Transformer in the context of nuclei segmentation, which thus can serve as an efficient solution for real clinical practice. Code is available at https://github.com/zhenqi-he/transnuseg.

Weiyu Guo, Zhaoshuo Li, Yongkui Yang et al.

Stereo depth estimation is of great interest for computer vision research. However, existing methods struggles to generalize and predict reliably in hazardous regions, such as large uniform regions. To overcome these limitations, we propose Context Enhanced Path (CEP). CEP improves the generalization and robustness against common failure cases in existing solutions by capturing the long-range global information. We construct our stereo depth estimation model, Context Enhanced Stereo Transformer (CSTR), by plugging CEP into the state-of-the-art stereo depth estimation method Stereo Transformer. CSTR is examined on distinct public datasets, such as Scene Flow, Middlebury-2014, KITTI-2015, and MPI-Sintel. We find CSTR outperforms prior approaches by a large margin. For example, in the zero-shot synthetic-to-real setting, CSTR outperforms the best competing approaches on Middlebury-2014 dataset by 11%. Our extensive experiments demonstrate that the long-range information is critical for stereo matching task and CEP successfully captures such information.

Zhaoshuo Li, Thomas Müller, Alex Evans et al.

Neural surface reconstruction has been shown to be powerful for recovering dense 3D surfaces via image-based neural rendering. However, current methods struggle to recover detailed structures of real-world scenes. To address the issue, we present Neuralangelo, which combines the representation power of multi-resolution 3D hash grids with neural surface rendering. Two key ingredients enable our approach: (1) numerical gradients for computing higher-order derivatives as a smoothing operation and (2) coarse-to-fine optimization on the hash grids controlling different levels of details. Even without auxiliary inputs such as depth, Neuralangelo can effectively recover dense 3D surface structures from multi-view images with fidelity significantly surpassing previous methods, enabling detailed large-scale scene reconstruction from RGB video captures.

Hao Ding, Lalithkumar Seenivasan, Hongchao Shu et al.

Large language model-based (LLM) agents are emerging as a powerful enabler of robust embodied intelligence due to their capability of planning complex action sequences. Sound planning ability is necessary for robust automation in many task domains, but especially in surgical automation. These agents rely on a highly detailed natural language representation of the scene. Thus, to leverage the emergent capabilities of LLM agents for surgical task planning, developing similarly powerful and robust perception algorithms is necessary to derive a detailed scene representation of the environment from visual input. Previous research has focused primarily on enabling LLM-based task planning while adopting simple yet severely limited perception solutions to meet the needs for bench-top experiments, but lacks the critical flexibility to scale to less constrained settings. In this work, we propose an alternate perception approach -- a digital twin (DT)-based machine perception approach that capitalizes on the convincing performance and out-of-the-box generalization of recent vision foundation models. Integrating our DT representation and LLM agent for planning with the dVRK platform, we develop an embodied intelligence system and evaluate its robustness in performing peg transfer and gauze retrieval tasks. Our approach shows strong task performance and generalizability to varied environmental settings. Despite a convincing performance, this work is merely a first step towards the integration of DT representations. Future studies are necessary for the realization of a comprehensive DT framework to improve the interpretability and generalizability of embodied intelligence in surgery.

Yiqing Shen, Pengfei Guo, Jingpu Wu et al.

The synergy of long-range dependencies from transformers and local representations of image content from convolutional neural networks (CNNs) has led to advanced architectures and increased performance for various medical image analysis tasks due to their complementary benefits. However, compared with CNNs, transformers require considerably more training data, due to a larger number of parameters and an absence of inductive bias. The need for increasingly large datasets continues to be problematic, particularly in the context of medical imaging, where both annotation efforts and data protection result in limited data availability. In this work, inspired by the human decision-making process of correlating new evidence with previously memorized experience, we propose a Memorizing Vision Transformer (MoViT) to alleviate the need for large-scale datasets to successfully train and deploy transformer-based architectures. MoViT leverages an external memory structure to cache history attention snapshots during the training stage. To prevent overfitting, we incorporate an innovative memory update scheme, attention temporal moving average, to update the stored external memories with the historical moving average. For inference speedup, we design a prototypical attention learning method to distill the external memory into smaller representative subsets. We evaluate our method on a public histology image dataset and an in-house MRI dataset, demonstrating that MoViT applied to varied medical image analysis tasks, can outperform vanilla transformer models across varied data regimes, especially in cases where only a small amount of annotated data is available. More importantly, MoViT can reach a competitive performance of ViT with only 3.0% of the training data.

Benjamin D. Killeen, Han Zhang, Jan Mangulabnan et al.

Surgical phase recognition (SPR) is a crucial element in the digital transformation of the modern operating theater. While SPR based on video sources is well-established, incorporation of interventional X-ray sequences has not yet been explored. This paper presents Pelphix, a first approach to SPR for X-ray-guided percutaneous pelvic fracture fixation, which models the procedure at four levels of granularity -- corridor, activity, view, and frame value -- simulating the pelvic fracture fixation workflow as a Markov process to provide fully annotated training data. Using added supervision from detection of bony corridors, tools, and anatomy, we learn image representations that are fed into a transformer model to regress surgical phases at the four granularity levels. Our approach demonstrates the feasibility of X-ray-based SPR, achieving an average accuracy of 93.8% on simulated sequences and 67.57% in cadaver across all granularity levels, with up to 88% accuracy for the target corridor in real data. This work constitutes the first step toward SPR for the X-ray domain, establishing an approach to categorizing phases in X-ray-guided surgery, simulating realistic image sequences to enable machine learning model development, and demonstrating that this approach is feasible for the analysis of real procedures. As X-ray-based SPR continues to mature, it will benefit procedures in orthopedic surgery, angiography, and interventional radiology by equipping intelligent surgical systems with situational awareness in the operating room.

Mitchell Pavlak, Nathan Drenkow, Nicholas Petrick et al.

To safely deploy deep learning-based computer vision models for computer-aided detection and diagnosis, we must ensure that they are robust and reliable. Towards that goal, algorithmic auditing has received substantial attention. To guide their audit procedures, existing methods rely on heuristic approaches or high-level objectives (e.g., non-discrimination in regards to protected attributes, such as sex, gender, or race). However, algorithms may show bias with respect to various attributes beyond the more obvious ones, and integrity issues related to these more subtle attributes can have serious consequences. To enable the generation of actionable, data-driven hypotheses which identify specific dataset attributes likely to induce model bias, we contribute a first technique for the rigorous, quantitative screening of medical image datasets. Drawing from literature in the causal inference and information theory domains, our procedure decomposes the risks associated with dataset attributes in terms of their detectability and utility (defined as the amount of information knowing the attribute gives about a task label). To demonstrate the effectiveness and sensitivity of our method, we develop a variety of datasets with synthetically inserted artifacts with different degrees of association to the target label that allow evaluation of inherited model biases via comparison of performance against true counterfactual examples. Using these datasets and results from hundreds of trained models, we show our screening method reliably identifies nearly imperceptible bias-inducing artifacts. Lastly, we apply our method to the natural attributes of a popular skin-lesion dataset and demonstrate its success. Our approach provides a means to perform more systematic algorithmic audits and guide future data collection efforts in pursuit of safer and more reliable models.

Zhaoshuo Li, Hongchao Shu, Ruixing Liang et al.

Purpose: Tracking the 3D motion of the surgical tool and the patient anatomy is a fundamental requirement for computer-assisted skull-base surgery. The estimated motion can be used both for intra-operative guidance and for downstream skill analysis. Recovering such motion solely from surgical videos is desirable, as it is compliant with current clinical workflows and instrumentation. Methods: We present Tracker of Anatomy and Tool (TAToo). TAToo jointly tracks the rigid 3D motion of patient skull and surgical drill from stereo microscopic videos. TAToo estimates motion via an iterative optimization process in an end-to-end differentiable form. For robust tracking performance, TAToo adopts a probabilistic formulation and enforces geometric constraints on the object level. Results: We validate TAToo on both simulation data, where ground truth motion is available, as well as on anthropomorphic phantom data, where optical tracking provides a strong baseline. We report sub-millimeter and millimeter inter-frame tracking accuracy for skull and drill, respectively, with rotation errors below 1°. We further illustrate how TAToo may be used in a surgical navigation setting. Conclusion: We present TAToo, which simultaneously tracks the surgical tool and the patient anatomy in skull-base surgery. TAToo directly predicts the motion from surgical videos, without the need of any markers. Our results show that the performance of TAToo compares favorably to competing approaches. Future work will include fine-tuning of our depth network to reach a 1 mm clinical accuracy goal desired for surgical applications in the skull base.

Cong Gao, Benjamin D. Killeen, Yicheng Hu et al.

Artificial intelligence (AI) now enables automated interpretation of medical images for clinical use. However, AI's potential use for interventional images (versus those involved in triage or diagnosis), such as for guidance during surgery, remains largely untapped. This is because surgical AI systems are currently trained using post hoc analysis of data collected during live surgeries, which has fundamental and practical limitations, including ethical considerations, expense, scalability, data integrity, and a lack of ground truth. Here, we demonstrate that creating realistic simulated images from human models is a viable alternative and complement to large-scale in situ data collection. We show that training AI image analysis models on realistically synthesized data, combined with contemporary domain generalization or adaptation techniques, results in models that on real data perform comparably to models trained on a precisely matched real data training set. Because synthetic generation of training data from human-based models scales easily, we find that our model transfer paradigm for X-ray image analysis, which we refer to as SyntheX, can even outperform real data-trained models due to the effectiveness of training on a larger dataset. We demonstrate the potential of SyntheX on three clinical tasks: Hip image analysis, surgical robotic tool detection, and COVID-19 lung lesion segmentation. SyntheX provides an opportunity to drastically accelerate the conception, design, and evaluation of intelligent systems for X-ray-based medicine. In addition, simulated image environments provide the opportunity to test novel instrumentation, design complementary surgical approaches, and envision novel techniques that improve outcomes, save time, or mitigate human error, freed from the ethical and practical considerations of live human data collection.

Hao Ding, Jie Ying Wu, Zhaoshuo Li et al.

Purpose: Vision-based robot tool segmentation plays a fundamental role in surgical robots and downstream tasks. CaRTS, based on a complementary causal model, has shown promising performance in unseen counterfactual surgical environments in the presence of smoke, blood, etc. However, CaRTS requires over 30 iterations of optimization to converge for a single image due to limited observability. Method: To address the above limitations, we take temporal relation into consideration and propose a temporal causal model for robot tool segmentation on video sequences. We design an architecture named Temporally Constrained CaRTS (TC-CaRTS). TC-CaRTS has three novel modules to complement CaRTS - temporal optimization pipeline, kinematics correction network, and spatial-temporal regularization. Results: Experiment results show that TC-CaRTS requires much fewer iterations to achieve the same or better performance as CaRTS. TC- CaRTS also has the same or better performance in different domains compared to CaRTS. All three modules are proven to be effective. Conclusion: We propose TC-CaRTS, which takes advantage of temporal constraints as additional observability. We show that TC-CaRTS outperforms prior work in the robot tool segmentation task with improved convergence speed on test datasets from different domains.

Jan Emily Mangulabnan, Roger D. Soberanis-Mukul, Timo Teufel et al.

Generating accurate 3D reconstructions from endoscopic video is a promising avenue for longitudinal radiation-free analysis of sinus anatomy and surgical outcomes. Several methods for monocular reconstruction have been proposed, yielding visually pleasant 3D anatomical structures by retrieving relative camera poses with structure-from-motion-type algorithms and fusion of monocular depth estimates. However, due to the complex properties of the underlying algorithms and endoscopic scenes, the reconstruction pipeline may perform poorly or fail unexpectedly. Further, acquiring medical data conveys additional challenges, presenting difficulties in quantitatively benchmarking these models, understanding failure cases, and identifying critical components that contribute to their precision. In this work, we perform a quantitative analysis of a self-supervised approach for sinus reconstruction using endoscopic sequences paired with optical tracking and high-resolution computed tomography acquired from nine ex-vivo specimens. Our results show that the generated reconstructions are in high agreement with the anatomy, yielding an average point-to-mesh error of 0.91 mm between reconstructions and CT segmentations. However, in a point-to-point matching scenario, relevant for endoscope tracking and navigation, we found average target registration errors of 6.58 mm. We identified that pose and depth estimation inaccuracies contribute equally to this error and that locally consistent sequences with shorter trajectories generate more accurate reconstructions. These results suggest that achieving global consistency between relative camera poses and estimated depths with the anatomy is essential. In doing so, we can ensure proper synergy between all components of the pipeline for improved reconstructions that will facilitate clinical application of this innovative technology.

Nathan Drenkow, Mathias Unberath

Object-centric representation learning offers the potential to overcome limitations of image-level representations by explicitly parsing image scenes into their constituent components. While image-level representations typically lack robustness to natural image corruptions, the robustness of object-centric methods remains largely untested. To address this gap, we present the RobustCLEVR benchmark dataset and evaluation framework. Our framework takes a novel approach to evaluating robustness by enabling the specification of causal dependencies in the image generation process grounded in expert knowledge and capable of producing a wide range of image corruptions unattainable in existing robustness evaluations. Using our framework, we define several causal models of the image corruption process which explicitly encode assumptions about the causal relationships and distributions of each corruption type. We generate dataset variants for each causal model on which we evaluate state-of-the-art object-centric methods. Overall, we find that object-centric methods are not inherently robust to image corruptions. Our causal evaluation approach exposes model sensitivities not observed using conventional evaluation processes, yielding greater insight into robustness differences across algorithms. Lastly, while conventional robustness evaluations view corruptions as out-of-distribution, we use our causal framework to show that even training on in-distribution image corruptions does not guarantee increased model robustness. This work provides a step towards more concrete and substantiated understanding of model performance and deterioration under complex corruption processes of the real-world.

Open-H-Embodiment Consortium, Nigel Nelson, Juo-Tung Chen et al.

Autonomous medical robots hold promise to improve patient outcomes, reduce provider workload, democratize access to care, and enable superhuman precision. However, autonomous medical robotics has been limited by a fundamental data problem: existing medical robotic datasets are small, single-embodiment, and rarely shared openly, restricting the development of foundation models that the field needs to advance. We introduce Open-H-Embodiment, the largest open dataset of medical robotic video with synchronized kinematics to date, spanning more than 49 institutions and multiple robotic platforms including the CMR Versius, Intuitive Surgical's da Vinci, da Vinci Research Kit (dVRK), Rob Surgical BiTrack, Virtual Incision's MIRA, Moon Surgical Maestro, and a variety of custom systems, spanning surgical manipulation, robotic ultrasound, and endoscopy procedures. We demonstrate the research enabled by this dataset through two foundation models. GR00T-H is the first open foundation vision-language-action model for medical robotics, which is the only evaluated model to achieve full end-to-end task completion on a structured suturing benchmark (25% of trials vs. 0% for all others) and achieves 64% average success across a 29-step ex vivo suturing sequence. We also train Cosmos-H-Surgical-Simulator, the first action-conditioned world model to enable multi-embodiment surgical simulation from a single checkpoint, spanning nine robotic platforms and supporting in silico policy evaluation and synthetic data generation for the medical domain. These results suggest that open, large-scale medical robot data collection can serve as critical infrastructure for the research community, enabling advances in robot learning, world modeling, and beyond.

Linda-Sophie Schneider, Mareike Thies, Christopher Syben et al.

We present a method for selecting valuable projections in computed tomography (CT) scans to enhance image reconstruction and diagnosis. The approach integrates two important factors, projection-based detectability and data completeness, into a single feed-forward neural network. The network evaluates the value of projections, processes them through a differentiable ranking function and makes the final selection using a straight-through estimator. Data completeness is ensured through the label provided during training. The approach eliminates the need for heuristically enforcing data completeness, which may exclude valuable projections. The method is evaluated on simulated data in a non-destructive testing scenario, where the aim is to maximize the reconstruction quality within a specified region of interest. We achieve comparable results to previous methods, laying the foundation for using reconstruction-based loss functions to learn the selection of projections.

Han Zhang, Yiqing Shen, Roger D. Soberanis-Mukul et al.

Developing embodied AI for intelligent surgical systems requires safe, controllable environments for continual learning and evaluation. However, safety regulations and operational constraints in operating rooms (ORs) limit embodied agents from freely perceiving and interacting in realistic settings. Digital twins provide high-fidelity, risk-free environments for exploration and training. How we may create photorealistic and dynamic digital representations of ORs that capture relevant spatial, visual, and behavioral complexity remains unclear. We introduce TwinOR, a framework for constructing photorealistic, dynamic digital twins of ORs for embodied AI research. The system reconstructs static geometry from pre-scan videos and continuously models human and equipment motion through multi-view perception of OR activities. The static and dynamic components are fused into an immersive 3D environment that supports controllable simulation and embodied exploration. The proposed framework reconstructs complete OR geometry with centimeter level accuracy while preserving dynamic interaction across surgical workflows, enabling realistic renderings and a virtual playground for embodied AI systems. In our experiments, TwinOR simulates stereo and monocular sensor streams for geometry understanding and visual localization tasks. Models such as FoundationStereo and ORB-SLAM3 on TwinOR-synthesized data achieve performance within their reported accuracy on real indoor datasets, demonstrating that TwinOR provides sensor-level realism sufficient for perception and localization challenges. By establishing a real-to-sim pipeline for constructing dynamic, photorealistic digital twins of OR environments, TwinOR enables the safe, scalable, and data-efficient development and benchmarking of embodied AI, ultimately accelerating the deployment of embodied AI from sim-to-real.

Catalina Gomez, Sue Min Cho, Shichang Ke et al.

Leveraging Artificial Intelligence (AI) in decision support systems has disproportionately focused on technological advancements, often overlooking the alignment between algorithmic outputs and human expectations. A human-centered perspective attempts to alleviate this concern by designing AI solutions for seamless integration with existing processes. Determining what information AI should provide to aid humans is vital, a concept underscored by explainable AI's efforts to justify AI predictions. However, how the information is presented, e.g., the sequence of recommendations and solicitation of interpretations, is equally crucial as complex interactions may emerge between humans and AI. While empirical studies have evaluated human-AI dynamics across domains, a common vocabulary for human-AI interaction protocols is lacking. To promote more deliberate consideration of interaction designs, we introduce a taxonomy of interaction patterns that delineate various modes of human-AI interactivity. We summarize the results of a systematic review of AI-assisted decision making literature and identify trends and opportunities in existing interactions across application domains from 105 articles. We find that current interactions are dominated by simplistic collaboration paradigms, leading to little support for truly interactive functionality. Our taxonomy offers a tool to understand interactivity with AI in decision-making and foster interaction designs for achieving clear communication, trustworthiness, and collaboration.

Anqi Feng, Dimitri Johnson, Grace R. Reilly et al.

Importance: Ultra-widefield fundus photography (UWF-FP) has shown utility in sickle cell retinopathy screening; however, image artifact may diminish quality and gradeability of images. Objective: To create an automated algorithm for UWF-FP artifact classification. Design: A neural network based automated artifact detection algorithm was designed to identify commonly encountered UWF-FP artifacts in a cross section of patient UWF-FP. A pre-trained ResNet-50 neural network was trained on a subset of the images and the classification accuracy, sensitivity, and specificity were quantified on the hold out test set. Setting: The study is based on patients from a tertiary care hospital site. Participants: There were 243 UWF-FP acquired from patients with sickle cell disease (SCD), and artifact labelling in the following categories was performed: Eyelash Present, Lower Eyelid Obstructing, Upper Eyelid Obstructing, Image Too Dark, Dark Artifact, and Image Not Centered. Results: Overall, the accuracy for each class was Eyelash Present at 83.7%, Lower Eyelid Obstructing at 83.7%, Upper Eyelid Obstructing at 98.0%, Image Too Dark at 77.6%, Dark Artifact at 93.9%, and Image Not Centered at 91.8%. Conclusions and Relevance: This automated algorithm shows promise in identifying common imaging artifacts on a subset of Optos UWF-FP in SCD patients. Further refinement is ongoing with the goal of improving efficiency of tele-retinal screening in sickle cell retinopathy (SCR) by providing a photographer real-time feedback as to the types of artifacts present, and the need for image re-acquisition. This algorithm also may have potential future applicability in other retinal diseases by improving quality and efficiency of image acquisition of UWF-FP.

Hongchao Shu, Ruixing Liang, Zhaoshuo Li et al.

Purpose: Digital twins are virtual interactive models of the real world, exhibiting identical behavior and properties. In surgical applications, computational analysis from digital twins can be used, for example, to enhance situational awareness. Methods: We present a digital twin framework for skull-base surgeries, named Twin-S, which can be integrated within various image-guided interventions seamlessly. Twin-S combines high-precision optical tracking and real-time simulation. We rely on rigorous calibration routines to ensure that the digital twin representation precisely mimics all real-world processes. Twin-S models and tracks the critical components of skull-base surgery, including the surgical tool, patient anatomy, and surgical camera. Significantly, Twin-S updates and reflects real-world drilling of the anatomical model in frame rate. Results: We extensively evaluate the accuracy of Twin-S, which achieves an average 1.39 mm error during the drilling process. We further illustrate how segmentation masks derived from the continuously updated digital twin can augment the surgical microscope view in a mixed reality setting, where bone requiring ablation is highlighted to provide surgeons additional situational awareness. Conclusion: We present Twin-S, a digital twin environment for skull-base surgery. Twin-S tracks and updates the virtual model in real-time given measurements from modern tracking technologies. Future research on complementing optical tracking with higher-precision vision-based approaches may further increase the accuracy of Twin-S.

Hao Ding, Yuqian Zhang, Tuxun Lu et al.

Surgical data science has seen rapid advancement due to the excellent performance of end-to-end deep neural networks (DNNs) for surgical video analysis. Despite their successes, end-to-end DNNs have been proven susceptible to even minor corruptions, substantially impairing the model's performance. This vulnerability has become a major concern for the translation of cutting-edge technology, especially for high-stakes decision-making in surgical data science. We introduce SegSTRONG-C, a benchmark and challenge in surgical data science dedicated, aiming to better understand model deterioration under unforeseen but plausible non-adversarial corruption and the capabilities of contemporary methods that seek to improve it. Through comprehensive baseline experiments and participating submissions from widespread community engagement, SegSTRONG-C reveals key themes for model failure and identifies promising directions for improving robustness. The performance of challenge winners, achieving an average 0.9394 DSC and 0.9301 NSD across the unreleased test sets with corruption types: bleeding, smoke, and low brightness, shows inspiring improvement of 0.1471 DSC and 0.2584 NSD in average comparing to strongest baseline methods with UNet architecture trained with AutoAugment. In conclusion, the SegSTRONG-C challenge has identified some practical approaches for enhancing model robustness, yet most approaches relied on conventional techniques that have known, and sometimes quite severe, limitations. Looking ahead, we advocate for expanding intellectual diversity and creativity in non-adversarial robustness beyond data augmentation or training scale, calling for new paradigms that enhance universal robustness to corruptions and may enable richer applications in surgical data science.

Hongchao Shu, Roger D. Soberanis-Mukul, Jiru Xu et al.

Accurate intra-operative localization of the bronchoscope tip relative to patient anatomy remains challenging due to respiratory motion, anatomical variability, and CT-to-body divergence that cause deformation and misalignment between intra-operative views and pre-operative CT. Existing vision-based methods often fail to generalize across domains and patients, leading to residual alignment errors. This work establishes a generalizable foundation for bronchoscopy navigation through a robust vision-based framework and a new synthetic benchmark dataset that enables standardized and reproducible evaluation. We propose a vision-based pose optimization framework for frame-wise 2D-3D registration between intra-operative endoscopic views and pre-operative CT anatomy. A fine-tuned modality- and domain-invariant encoder enables direct similarity computation between real endoscopic RGB frames and CT-rendered depth maps, while a differentiable rendering module iteratively refines camera poses through depth consistency. To enhance reproducibility, we introduce the first public synthetic benchmark dataset for bronchoscopy navigation, addressing the lack of paired CT-endoscopy data. Trained exclusively on synthetic data distinct from the benchmark, our model achieves an average translational error of 2.65 mm and a rotational error of 0.19 rad, demonstrating accurate and stable localization. Qualitative results on real patient data further confirm strong cross-domain generalization, achieving consistent frame-wise 2D-3D alignment without domain-specific adaptation. Overall, the proposed framework achieves robust, domain-invariant localization through iterative vision-based optimization, while the new benchmark provides a foundation for standardized progress in vision-based bronchoscopy navigation.

Aiza Maksutova, Lalithkumar Seenivasan, Hao Ding et al.

Surgical action automation has progressed rapidly toward achieving surgeon-like dexterous control, driven primarily by advances in learning from demonstration and vision-language-action models. While these have demonstrated success in table-top experiments, translating them to clinical deployment remains challenging: current methods offer limited predictability on where instruments will interact on tissue surfaces and lack explicit conditioning inputs to enforce tool-action-specific safe interaction regions. Addressing this gap, we introduce AffordTissue, a multimodal framework for predicting tool-action specific tissue affordance regions as dense heatmaps during cholecystectomy. Our approach combines a temporal vision encoder capturing tool motion and tissue dynamics across multiple viewpoints, language conditioning enabling generalization across diverse instrument-action pairs, and a DiT-style decoder for dense affordance prediction. We establish the first tissue affordance benchmark by curating and annotating 15,638 video clips across 103 cholecystectomy procedures, covering six unique tool-action pairs involving four instruments (hook, grasper, scissors, clipper) and their associated tasks: dissection, grasping, clipping, and cutting. Experiments demonstrate substantial improvement over vision-language model baselines (20.6 px ASSD vs. 60.2 px for Molmo-VLM), showing that our task-specific architecture outperforms large-scale foundation models for dense surgical affordance prediction. By predicting tool-action specific tissue affordance regions, AffordTissue provides explicit spatial reasoning for safe surgical automation, potentially unlocking explicit policy guidance toward appropriate tissue regions and early safe stop when instruments deviate outside predicted safe zones.

Jan Emily Mangulabnan, Akshat Chauhan, Laura Fleig et al.

In endoscopic surgery, surgeons continuously locate the endoscopic view relative to the anatomy by interpreting the evolving visual appearance of the intraoperative scene in the context of their prior knowledge. Vision-based navigation systems seek to replicate this capability by recovering camera pose directly from endoscopic video, but most approaches do not embody the same principles of reasoning about new frames that makes surgeons successful. Instead, they remain grounded in feature matching and geometric optimization over keyframes, an approach that has been shown to degrade under the challenging conditions of endoscopic imaging like low texture and rapid illumination changes. Here, we pursue an alternative approach and investigate a policy-based formulation of endoscopic camera pose recovery that seeks to imitate experts in estimating trajectories conditioned on the previous camera state. Our approach directly predicts short-horizon relative motions without maintaining an explicit geometric representation at inference time. It thus addresses, by design, some of the notorious challenges of geometry-based approaches, such as brittle correspondence matching, instability in texture-sparse regions, and limited pose coverage due to reconstruction failure. We evaluate the proposed formulation on cadaveric sinus endoscopy. Under oracle state conditioning, we compare short-horizon motion prediction quality to geometric baselines achieving lowest mean translation error and competitive rotational accuracy. We analyze robustness by grouping prediction windows according to texture richness and illumination change indicating reduced sensitivity to low-texture conditions. These findings suggest that a learned motion policy offers a viable alternative formulation for endoscopic camera pose recovery.

Florence Klitzner, Blanca Inigo, Benjamin D. Killeen et al.

Imitation learning-based robot control policies are enjoying renewed interest in video-based robotics. However, it remains unclear whether this approach applies to X-ray-guided procedures, such as spine instrumentation. This is because interpretation of multi-view X-rays is complex. We examine opportunities and challenges for imitation policy learning in bi-plane-guided cannula insertion. We develop an in silico sandbox for scalable, automated simulation of X-ray-guided spine procedures with a high degree of realism. We curate a dataset of correct trajectories and corresponding bi-planar X-ray sequences that emulate the stepwise alignment of providers. We then train imitation learning policies for planning and open-loop control that iteratively align a cannula solely based on visual information. This precisely controlled setup offers insights into limitations and capabilities of this method. Our policy succeeded on the first attempt in 68.5% of cases, maintaining safe intra-pedicular trajectories across diverse vertebral levels. The policy generalized to complex anatomy, including fractures, and remained robust to varied initializations. Rollouts on real bi-planar X-rays further suggest that the model can produce plausible trajectories, despite training exclusively in simulation. While these preliminary results are promising, we also identify limitations, especially in entry point precision. Full closed-look control will require additional considerations around how to provide sufficiently frequent feedback. With more robust priors and domain knowledge, such models may provide a foundation for future efforts toward lightweight and CT-free robotic intra-operative spinal navigation.

Dominik Schneider, Lalithkumar Seenivasan, Sampath Rapuri et al.

Purpose: Curating large-scale datasets of operating room (OR) workflow, encompassing rare, safety-critical, or atypical events, remains operationally and ethically challenging. This data bottleneck complicates the development of ambient intelligence for detecting, understanding, and mitigating rare or safety-critical events in the OR. Methods: This work presents an OR video diffusion framework that enables controlled synthesis of rare and safety-critical events. The framework integrates a geometric abstraction module, a conditioning module, and a fine-tuned diffusion model to first transform OR scenes into abstract geometric representations, then condition the synthesis process, and finally generate realistic OR event videos. Using this framework, we also curate a synthetic dataset to train and validate AI models for detecting near-misses of sterile-field violations. Results: In synthesizing routine OR events, our method outperforms off-the-shelf video diffusion baselines, achieving lower FVD/LPIPS and higher SSIM/PSNR in both in- and out-of-domain datasets. Through qualitative results, we illustrate its ability for controlled video synthesis of counterfactual events. An AI model trained and validated on the generated synthetic data achieved a RECALL of 70.13% in detecting near safety-critical events. Finally, we conduct an ablation study to quantify performance gains from key design choices. Conclusion: Our solution enables controlled synthesis of routine and rare OR events from abstract geometric representations. Beyond demonstrating its capability to generate rare and safety-critical scenarios, we show its potential to support the development of ambient intelligence models.

Hongchao Shu, Lalithkumar Seenivasan, Mingxu Liu et al.

Arthroscopic procedures can greatly benefit from navigation systems that enhance spatial awareness, depth perception, and field of view. However, existing optical tracking solutions impose strict workspace constraints and disrupt surgical workflow. Vision-based alternatives, though less invasive, often rely solely on the monocular arthroscope camera, making them prone to drift, scale ambiguity, and sensitivity to rapid motion or occlusion. We propose DualVision ArthroNav, a multi-camera arthroscopy navigation system that integrates an external camera rigidly mounted on the arthroscope. The external camera provides stable visual odometry and absolute localization, while the monocular arthroscope video enables dense scene reconstruction. By combining these complementary views, our system resolves the scale ambiguity and long-term drift inherent in monocular SLAM and ensures robust relocalization. Experiments demonstrate that our system effectively compensates for calibration errors, achieving an average absolute trajectory error of 1.09 mm. The reconstructed scenes reach an average target registration error of 2.16 mm, with high visual fidelity (SSIM = 0.69, PSNR = 22.19). These results indicate that our system provides a practical and cost-efficient solution for arthroscopic navigation, bridging the gap between optical tracking and purely vision-based systems, and paving the way toward clinically deployable, fully vision-based arthroscopic guidance.

Yiqing Shen, Chenxiao Fan, Chenjia Li et al.

The goal of text-to-video retrieval is to search large databases for relevant videos based on text queries. Existing methods have progressed to handling explicit queries where the visual content of interest is described explicitly; however, they fail with implicit queries where identifying videos relevant to the query requires reasoning. We introduce reasoning text-to-video retrieval, a paradigm that extends traditional retrieval to process implicit queries through reasoning while providing object-level grounding masks that identify which entities satisfy the query conditions. Instead of relying on vision-language models directly, we propose representing video content as digital twins, i.e., structured scene representations that decompose salient objects through specialist vision models. This approach is beneficial because it enables large language models to reason directly over long-horizon video content without visual token compression. Specifically, our two-stage framework first performs compositional alignment between decomposed sub-queries and digital twin representations for candidate identification, then applies large language model-based reasoning with just-in-time refinement that invokes additional specialist models to address information gaps. We construct a benchmark of 447 manually created implicit queries with 135 videos (ReasonT2VBench-135) and another more challenging version of 1000 videos (ReasonT2VBench-1000). Our method achieves 81.2% R@1 on ReasonT2VBench-135, outperforming the strongest baseline by greater than 50 percentage points, and maintains 81.7% R@1 on the extended configuration while establishing state-of-the-art results in three conventional benchmarks (MSR-VTT, MSVD, and VATEX).

Yiqing Shen, Mathias Unberath

Reasoning segmentation enables open-set object segmentation via implicit text queries, therefore serving as a foundation for embodied agents that should operate autonomously in real-world environments. However, existing methods for reasoning segmentation require multimodal large language models with billions of parameters that exceed the computational capabilities of edge devices that typically deploy the embodied AI systems. Distillation offers a pathway to compress these models while preserving their capabilities. Yet, existing distillation approaches fail to transfer the multi-step reasoning capabilities that reasoning segmentation demands, as they focus on matching output predictions and intermediate features rather than preserving reasoning chains. The emerging paradigm of reasoning over digital twin representations presents an opportunity for more effective distillation by re-framing the problem. Consequently, we propose FastReasonSeg, which employs digital twin representations that decouple perception from reasoning to enable more effective distillation. Our distillation scheme first relies on supervised fine-tuning on teacher-generated reasoning chains. Then it is followed by reinforcement fine-tuning with joint rewards evaluating both segmentation accuracy and reasoning quality alignment. Experiments on two video (JiTBench, RVTBench) and two image benchmarks (ReasonSeg, LLM-Seg40K) demonstrate that our FastReasonSeg achieves state-of-the-art reasoning segmentation performance. Moreover, the distilled 0.6B variant outperforms models with 20 times more parameters while achieving 7.79 FPS throughput with only 2.1GB memory consumption. This efficiency enables deployment in resource-constrained environments to enable real-time reasoning segmentation.

Yiqing Shen, Mathias Unberath

Visual reasoning may require models to interpret images and videos and respond to implicit text queries across diverse output formats, from pixel-level segmentation masks to natural language descriptions. Existing approaches rely on supervised fine-tuning with task-specific architectures. For example, reasoning segmentation, grounding, summarization, and visual question answering each demand distinct model designs and training, preventing unified solutions and limiting cross-task and cross-modality generalization. Hence, we propose DT-R1, a reinforcement learning framework that trains large language models to construct digital twin representations of complex multi-modal visual inputs and then reason over these high-level representations as a unified approach to visual reasoning. Specifically, we train DT-R1 using GRPO with a novel reward that validates both structural integrity and output accuracy. Evaluations in six visual reasoning benchmarks, covering two modalities and four task types, demonstrate that DT-R1 consistently achieves improvements over state-of-the-art task-specific models. DT-R1 opens a new direction where visual reasoning emerges from reinforcement learning with digital twin representations.

Xinyuan Shao, Yiqing Shen, Mathias Unberath

Segment Anything Models (SAMs) have gained increasing attention in medical image analysis due to their zero-shot generalization capability in segmenting objects of unseen classes and domains when provided with appropriate user prompts. Addressing this performance gap is important to fully leverage the pre-trained weights of SAMs, particularly in the domain of volumetric medical image segmentation, where accuracy is important but well-annotated 3D medical data for fine-tuning is limited. In this work, we investigate whether introducing the memory mechanism as a plug-in, specifically the ability to memorize and recall internal representations of past inputs, can improve the performance of SAM with limited computation cost. To this end, we propose Memorizing SAM, a novel 3D SAM architecture incorporating a memory Transformer as a plug-in. Unlike conventional memorizing Transformers that save the internal representation during training or inference, our Memorizing SAM utilizes existing highly accurate internal representation as the memory source to ensure the quality of memory. We evaluate the performance of Memorizing SAM in 33 categories from the TotalSegmentator dataset, which indicates that Memorizing SAM can outperform state-of-the-art 3D SAM variant i.e., FastSAM3D with an average Dice increase of 11.36% at the cost of only 4.38 millisecond increase in inference time. The source code is publicly available at https://github.com/swedfr/memorizingSAM

Benjamin D. Killeen, Liam J. Wang, Blanca Inigo et al.

Language promptable X-ray image segmentation would enable greater flexibility for human-in-the-loop workflows in diagnostic and interventional precision medicine. Prior efforts have contributed task-specific models capable of solving problems within a narrow scope, but expanding to broader use requires additional data, annotations, and training time. Recently, language-aligned foundation models (LFMs) -- machine learning models trained on large amounts of highly variable image and text data thus enabling broad applicability -- have emerged as promising tools for automated image analysis. Existing foundation models for medical image analysis focus on scenarios and modalities where large, richly annotated datasets are available. However, the X-ray imaging modality features highly variable image appearance and applications, from diagnostic chest X-rays to interventional fluoroscopy, with varying availability of data. To pave the way toward an LFM for comprehensive and language-aligned analysis of arbitrary medical X-ray images, we introduce FluoroSAM, a language-promptable variant of the Segment Anything Model, trained from scratch on 3M synthetic X-ray images from a wide variety of human anatomies, imaging geometries, and viewing angles. These include pseudo-ground truth masks for 128 organ types and 464 tools with associated text descriptions. FluoroSAM is capable of segmenting myriad anatomical structures and tools based on natural language prompts, thanks to the novel incorporation of vector quantization (VQ) of text embeddings in the training process. We demonstrate FluoroSAM's performance quantitatively on real X-ray images and showcase on several applications how FluoroSAM is a key enabler for rich human-machine interaction in the X-ray image acquisition and analysis context. Code is available at https://github.com/arcadelab/fluorosam.

Yiqing Shen, Jingxing Li, Xinyuan Shao et al.

Segment anything models (SAMs) are gaining attention for their zero-shot generalization capability in segmenting objects of unseen classes and in unseen domains when properly prompted. Interactivity is a key strength of SAMs, allowing users to iteratively provide prompts that specify objects of interest to refine outputs. However, to realize the interactive use of SAMs for 3D medical imaging tasks, rapid inference times are necessary. High memory requirements and long processing delays remain constraints that hinder the adoption of SAMs for this purpose. Specifically, while 2D SAMs applied to 3D volumes contend with repetitive computation to process all slices independently, 3D SAMs suffer from an exponential increase in model parameters and FLOPS. To address these challenges, we present FastSAM3D which accelerates SAM inference to 8 milliseconds per 128*128*128 3D volumetric image on an NVIDIA A100 GPU. This speedup is accomplished through 1) a novel layer-wise progressive distillation scheme that enables knowledge transfer from a complex 12-layer ViT-B to a lightweight 6-layer ViT-Tiny variant encoder without training from scratch; and 2) a novel 3D sparse flash attention to replace vanilla attention operators, substantially reducing memory needs and improving parallelization. Experiments on three diverse datasets reveal that FastSAM3D achieves a remarkable speedup of 527.38x compared to 2D SAMs and 8.75x compared to 3D SAMs on the same volumes without significant performance decline. Thus, FastSAM3D opens the door for low-cost truly interactive SAM-based 3D medical imaging segmentation with commonly used GPU hardware. Code is available at https://github.com/arcadelab/FastSAM3D.

Xingtong Liu, Zhaoshuo Li, Masaru Ishii et al.

In endoscopy, many applications (e.g., surgical navigation) would benefit from a real-time method that can simultaneously track the endoscope and reconstruct the dense 3D geometry of the observed anatomy from a monocular endoscopic video. To this end, we develop a Simultaneous Localization and Mapping system by combining the learning-based appearance and optimizable geometry priors and factor graph optimization. The appearance and geometry priors are explicitly learned in an end-to-end differentiable training pipeline to master the task of pair-wise image alignment, one of the core components of the SLAM system. In our experiments, the proposed SLAM system is shown to robustly handle the challenges of texture scarceness and illumination variation that are commonly seen in endoscopy. The system generalizes well to unseen endoscopes and subjects and performs favorably compared with a state-of-the-art feature-based SLAM system. The code repository is available at https://github.com/lppllppl920/SAGE-SLAM.git.

Xingtong Liu, Maia Stiber, Jindan Huang et al.

Reconstructing accurate 3D surface models of sinus anatomy directly from an endoscopic video is a promising avenue for cross-sectional and longitudinal analysis to better understand the relationship between sinus anatomy and surgical outcomes. We present a patient-specific, learning-based method for 3D reconstruction of sinus surface anatomy directly and only from endoscopic videos. We demonstrate the effectiveness and accuracy of our method on in and ex vivo data where we compare to sparse reconstructions from Structure from Motion, dense reconstruction from COLMAP, and ground truth anatomy from CT. Our textured reconstructions are watertight and enable measurement of clinically relevant parameters in good agreement with CT. The source code is available at https://github.com/lppllppl920/DenseReconstruction-Pytorch.

Jie Ying Wu, Peter Kazanzides, Mathias Unberath

Purpose Surgical simulations play an increasingly important role in surgeon education and developing algorithms that enable robots to perform surgical subtasks. To model anatomy, Finite Element Method (FEM) simulations have been held as the gold standard for calculating accurate soft-tissue deformation. Unfortunately, their accuracy is highly dependent on the simulation parameters, which can be difficult to obtain. Methods In this work, we investigate how live data acquired during any robotic endoscopic surgical procedure may be used to correct for inaccurate FEM simulation results. Since FEMs are calculated from initial parameters and cannot directly incorporate observations, we propose to add a correction factor that accounts for the discrepancy between simulation and observations. We train a network to predict this correction factor. Results To evaluate our method, we use an open-source da Vinci Surgical System to probe a soft-tissue phantom and replay the interaction in simulation. We train the network to correct for the difference between the predicted mesh position and the measured point cloud. This results in 15-30% improvement in the mean distance, demonstrating the effectiveness of our approach across a large range of simulation parameters. Conclusion We show a first step towards a framework that synergistically combines the benefits of model-based simulation and real-time observations. It corrects discrepancies between simulation and the scene that results from inaccurate modeling parameters. This can provide a more accurate simulation environment for surgeons and better data with which to train algorithms.

Xingtong Liu, Yiping Zheng, Benjamin Killeen et al.

High-quality 3D reconstructions from endoscopy video play an important role in many clinical applications, including surgical navigation where they enable direct video-CT registration. While many methods exist for general multi-view 3D reconstruction, these methods often fail to deliver satisfactory performance on endoscopic video. Part of the reason is that local descriptors that establish pair-wise point correspondences, and thus drive reconstruction, struggle when confronted with the texture-scarce surface of anatomy. Learning-based dense descriptors usually have larger receptive fields enabling the encoding of global information, which can be used to disambiguate matches. In this work, we present an effective self-supervised training scheme and novel loss design for dense descriptor learning. In direct comparison to recent local and dense descriptors on an in-house sinus endoscopy dataset, we demonstrate that our proposed dense descriptor can generalize to unseen patients and scopes, thereby largely improving the performance of Structure from Motion (SfM) in terms of model density and completeness. We also evaluate our method on a public dense optical flow dataset and a small-scale SfM public dataset to further demonstrate the effectiveness and generality of our method. The source code is available at https://github.com/lppllppl920/DenseDescriptorLearning-Pytorch.

Laura Fink, Sing Chun Lee, Jie Ying Wu et al.

With the increasing computational power of today's workstations, real-time physically-based rendering is within reach, rapidly gaining attention across a variety of domains. These have expeditiously applied to medicine, where it is a powerful tool for intuitive 3D data visualization. Embedded devices such as optical see-through head-mounted displays (OST HMDs) have been a trend for medical augmented reality. However, leveraging the obvious benefits of physically-based rendering remains challenging on these devices because of limited computational power, memory usage, and power consumption. We navigate the compromise between device limitations and image quality to achieve reasonable rendering results by introducing a novel light field that can be sampled in real-time on embedded devices. We demonstrate its applications in medicine and discuss limitations of the proposed method. An open-source version of this project is available at https://github.com/lorafib/LumiPath which provides full insight on implementation and exemplary demonstrational material.

Xingtong Liu, Ayushi Sinha, Masaru Ishii et al.

We present a self-supervised approach to training convolutional neural networks for dense depth estimation from monocular endoscopy data without a priori modeling of anatomy or shading. Our method only requires monocular endoscopic videos and a multi-view stereo method, e.g., structure from motion, to supervise learning in a sparse manner. Consequently, our method requires neither manual labeling nor patient computed tomography (CT) scan in the training and application phases. In a cross-patient experiment using CT scans as groundtruth, the proposed method achieved submillimeter mean residual error. In a comparison study to recent self-supervised depth estimation methods designed for natural video on in vivo sinus endoscopy data, we demonstrate that the proposed approach outperforms the previous methods by a large margin. The source code for this work is publicly available online at https://github.com/lppllppl920/EndoscopyDepthEstimation-Pytorch.

Yiqing Shen, Bohan Liu, Chenjia Li et al.

Reasoning segmentation (RS) aims to identify and segment objects of interest based on implicit text queries. As such, RS is a catalyst for embodied AI agents, enabling them to interpret high-level commands without requiring explicit step-by-step guidance. However, current RS approaches rely heavily on the visual perception capabilities of multimodal large language models (LLMs), leading to several major limitations. First, they struggle with queries that require multiple steps of reasoning or those that involve complex spatial/temporal relationships. Second, they necessitate LLM fine-tuning, which may require frequent updates to maintain compatibility with contemporary LLMs and may increase risks of catastrophic forgetting during fine-tuning. Finally, being primarily designed for static images or offline video processing, they scale poorly to online video data. To address these limitations, we propose an agent framework that disentangles perception and reasoning for online video RS without LLM fine-tuning. Our innovation is the introduction of a just-in-time digital twin concept, where -- given an implicit query -- a LLM plans the construction of a low-level scene representation from high-level video using specialist vision models. We refer to this approach to creating a digital twin as "just-in-time" because the LLM planner will anticipate the need for specific information and only request this limited subset instead of always evaluating every specialist model. The LLM then performs reasoning on this digital twin representation to identify target objects. To evaluate our approach, we introduce a new comprehensive video reasoning segmentation benchmark comprising 200 videos with 895 implicit text queries. The benchmark spans three reasoning categories (semantic, spatial, and temporal) with three different reasoning chain complexity.

Yiqing Shen, Chenjia Li, Fei Xiong et al.

Reasoning Segmentation (RS) aims to delineate objects based on implicit text queries, the interpretation of which requires reasoning and knowledge integration. Unlike the traditional formulation of segmentation problems that relies on fixed semantic categories or explicit prompting, RS bridges the gap between visual perception and human-like reasoning capabilities, facilitating more intuitive human-AI interaction through natural language. Our work presents the first comprehensive survey of RS for image and video processing, examining 26 state-of-the-art methods together with a review of the corresponding evaluation metrics, as well as 29 datasets and benchmarks. We also explore existing applications of RS across diverse domains and identify their potential extensions. Finally, we identify current research gaps and highlight promising future directions.

Yiqing Shen, Chenjia Li, Chenxiao Fan et al.

Visual reasoning, the capability to interpret visual input in response to implicit text query through multi-step reasoning, remains a challenge for deep learning models due to the lack of relevant benchmarks. Previous work in visual reasoning has primarily focused on reasoning segmentation, where models aim to segment objects based on implicit text queries. This paper introduces reasoning visual tasks (RVTs), a unified formulation that extends beyond traditional video reasoning segmentation to a diverse family of visual language reasoning problems, which can therefore accommodate multiple output formats including bounding boxes, natural language descriptions, and question-answer pairs. Correspondingly, we identify the limitations in current benchmark construction methods that rely solely on large language models (LLMs), which inadequately capture complex spatial-temporal relationships and multi-step reasoning chains in video due to their reliance on token representation, resulting in benchmarks with artificially limited reasoning complexity. To address this limitation, we propose a novel automated RVT benchmark construction pipeline that leverages digital twin (DT) representations as structured intermediaries between perception and the generation of implicit text queries. Based on this method, we construct RVTBench, a RVT benchmark containing 3,896 queries of over 1.2 million tokens across four types of RVT (segmentation, grounding, VQA and summary), three reasoning categories (semantic, spatial, and temporal), and four increasing difficulty levels, derived from 200 video sequences. Finally, we propose RVTagent, an agent framework for RVT that allows for zero-shot generalization across various types of RVT without task-specific fine-tuning.

Yiqing Shen, Hao Ding, Lalithkumar Seenivasan et al.

Current foundation models (FMs) rely on token representations that directly fragment continuous real-world multimodal data into discrete tokens. They limit FMs to learning real-world knowledge and relationships purely through statistical correlation rather than leveraging explicit domain knowledge. Consequently, current FMs struggle with maintaining semantic coherence across modalities, capturing fine-grained spatial-temporal dynamics, and performing causal reasoning. These limitations cannot be overcome by simply scaling up model size or expanding datasets. This position paper argues that the machine learning community should consider digital twin (DT) representations, which are outcome-driven digital representations that serve as building blocks for creating virtual replicas of physical processes, as an alternative to the token representation for building FMs. Finally, we discuss how DT representations can address these challenges by providing physically grounded representations that explicitly encode domain knowledge and preserve the continuous nature of real-world processes.

Kanyifeechukwu J. Oguine, Roger D. Soberanis-Mukul, Nathan Drenkow et al.

Purpose: Accurate tool segmentation is essential in computer-aided procedures. However, this task conveys challenges due to artifacts' presence and the limited training data in medical scenarios. Methods that generalize to unseen data represent an interesting venue, where zero-shot segmentation presents an option to account for data limitation. Initial exploratory works with the Segment Anything Model (SAM) show that bounding-box-based prompting presents notable zero-short generalization. However, point-based prompting leads to a degraded performance that further deteriorates under image corruption. We argue that SAM drastically over-segment images with high corruption levels, resulting in degraded performance when only a single segmentation mask is considered, while the combination of the masks overlapping the object of interest generates an accurate prediction. Method: We use SAM to generate the over-segmented prediction of endoscopic frames. Then, we employ the ground-truth tool mask to analyze the results of SAM when the best single mask is selected as prediction and when all the individual masks overlapping the object of interest are combined to obtain the final predicted mask. We analyze the Endovis18 and Endovis17 instrument segmentation datasets using synthetic corruptions of various strengths and an In-House dataset featuring counterfactually created real-world corruptions. Results: Combining the over-segmented masks contributes to improvements in the IoU. Furthermore, selecting the best single segmentation presents a competitive IoU score for clean images. Conclusions: Combined SAM predictions present improved results and robustness up to a certain corruption level. However, appropriate prompting strategies are fundamental for implementing these models in the medical domain.

Nathan Drenkow, Mitchell Pavlak, Keith Harrigian et al.

Artificial Intelligence (AI) is now firmly at the center of evidence-based medicine. Despite many success stories that edge the path of AI's rise in healthcare, there are comparably many reports of significant shortcomings and unexpected behavior of AI in deployment. A major reason for these limitations is AI's reliance on association-based learning, where non-representative machine learning datasets can amplify latent bias during training and/or hide it during testing. To unlock new tools capable of foreseeing and preventing such AI bias issues, we present G-AUDIT. Generalized Attribute Utility and Detectability-Induced bias Testing (G-AUDIT) for datasets is a modality-agnostic dataset auditing framework that allows for generating targeted hypotheses about sources of bias in training or testing data. Our method examines the relationship between task-level annotations (commonly referred to as ``labels'') and data properties including patient attributes (e.g., age, sex) and environment/acquisition characteristics (e.g., clinical site, imaging protocols). G-AUDIT quantifies the extent to which the observed data attributes pose a risk for shortcut learning, or in the case of testing data, might hide predictions made based on spurious associations. We demonstrate the broad applicability of our method by analyzing large-scale medical datasets for three distinct modalities and machine learning tasks: skin lesion classification in images, stigmatizing language classification in Electronic Health Records (EHR), and mortality prediction for ICU tabular data. In each setting, G-AUDIT successfully identifies subtle biases commonly overlooked by traditional qualitative methods, underscoring its practical value in exposing dataset-level risks and supporting the downstream development of reliable AI systems.

Hao Ding, Zhongpai Gao, Benjamin Planche et al.

Surgical phase recognition (SPR) is crucial for applications in workflow optimization, performance evaluation, and real-time intervention guidance. However, current deep learning models often struggle with fragmented predictions, failing to capture the sequential nature of surgical workflows. We propose the Neural Finite-State Machine (NFSM), a novel approach that enforces temporal coherence by integrating classical state-transition priors with modern neural networks. NFSM leverages learnable global state embeddings as unique phase identifiers and dynamic transition tables to model phase-to-phase progressions. Additionally, a future phase forecasting mechanism employs repeated frame padding to anticipate upcoming transitions. Implemented as a plug-and-play module, NFSM can be integrated into existing SPR pipelines without changing their core architectures. We demonstrate state-of-the-art performance across multiple benchmarks, including a significant improvement on the BernBypass70 dataset - raising video-level accuracy by 0.9 points and phase-level precision, recall, F1-score, and mAP by 3.8, 3.1, 3.3, and 4.1, respectively. Ablation studies confirm each component's effectiveness and the module's adaptability to various architectures. By unifying finite-state principles with deep learning, NFSM offers a robust path toward consistent, long-term surgical video analysis.

Nathan Drenkow, William Paul, Chris Ribaudo et al.

Robustness audits of deep neural networks (DNN) provide a means to uncover model sensitivities to the challenging real-world imaging conditions that significantly degrade DNN performance in-the-wild. Such conditions are often the result of multiple interacting factors inherent to the environment, sensor, or processing pipeline and may lead to complex image distortions that are not easily categorized. When robustness audits are limited to a set of isolated imaging effects or distortions, the results cannot be (easily) transferred to real-world conditions where image corruptions may be more complex or nuanced. To address this challenge, we present a new alternative robustness auditing method that uses causal inference to measure DNN sensitivities to the factors of the imaging process that cause complex distortions. Our approach uses causal models to explicitly encode assumptions about the domain-relevant factors and their interactions. Then, through extensive experiments on natural and rendered images across multiple vision tasks, we show that our approach reliably estimates causal effects of each factor on DNN performance using only observational domain data. These causal effects directly tie DNN sensitivities to observable properties of the imaging pipeline in the domain of interest towards reducing the risk of unexpected DNN failures when deployed in that domain.

Alejandra Perez, Han Zhang, Yu-Chun Ku et al.

The operating room (OR) is a complex environment where optimizing workflows is critical to reduce costs and improve patient outcomes. While computer vision approaches for automatic recognition of perioperative events can identify bottlenecks for OR optimization, privacy concerns limit the use of OR videos for automated event detection. We propose a two-stage pipeline for privacy-preserving OR video analysis and event detection. First, we leverage vision foundation models for depth estimation and semantic segmentation to generate de-identified Digital Twins (DT) of the OR from conventional RGB videos. Second, we employ the SafeOR model, a fused two-stream approach that processes segmentation masks and depth maps for OR event detection. Evaluation on an internal dataset of 38 simulated surgical trials with five event classes shows that our DT-based approach achieves performance on par with -- and sometimes better than -- raw RGB video-based models for OR event detection. Digital Twins enable privacy-preserving OR workflow analysis, facilitating the sharing of de-identified data across institutions and potentially enhancing model generalizability by mitigating domain-specific appearance differences.

Benjamin D. Killeen, Anushri Suresh, Catalina Gomez et al.

Natural language offers a convenient, flexible interface for controlling robotic C-arm X-ray systems, making advanced functionality and controls accessible. However, enabling language interfaces requires specialized AI models that interpret X-ray images to create a semantic representation for reasoning. The fixed outputs of such AI models limit the functionality of language controls. Incorporating flexible, language-aligned AI models prompted through language enables more versatile interfaces for diverse tasks and procedures. Using a language-aligned foundation model for X-ray image segmentation, our system continually updates a patient digital twin based on sparse reconstructions of desired anatomical structures. This supports autonomous capabilities such as visualization, patient-specific viewfinding, and automatic collimation from novel viewpoints, enabling commands 'Focus in on the lower lumbar vertebrae.' In a cadaver study, users visualized, localized, and collimated structures across the torso using verbal commands, achieving 84% end-to-end success. Post hoc analysis of randomly oriented images showed our patient digital twin could localize 35 commonly requested structures to within 51.68 mm, enabling localization and isolation from arbitrary orientations. Our results demonstrate how intelligent robotic X-ray systems can incorporate physicians' expressed intent directly. While existing foundation models for intra-operative X-ray analysis exhibit failure modes, as they improve, they can facilitate highly flexible, intelligent robotic C-arms.

Jan Emily Mangulabnan, Roger D. Soberanis-Mukul, Timo Teufel et al.

Purpose: Preoperative imaging plays a pivotal role in sinus surgery where CTs offer patient-specific insights of complex anatomy, enabling real-time intraoperative navigation to complement endoscopy imaging. However, surgery elicits anatomical changes not represented in the preoperative model, generating an inaccurate basis for navigation during surgery progression. Methods: We propose a first vision-based approach to update the preoperative 3D anatomical model leveraging intraoperative endoscopic video for navigated sinus surgery where relative camera poses are known. We rely on comparisons of intraoperative monocular depth estimates and preoperative depth renders to identify modified regions. The new depths are integrated in these regions through volumetric fusion in a truncated signed distance function representation to generate an intraoperative 3D model that reflects tissue manipulation. Results: We quantitatively evaluate our approach by sequentially updating models for a five-step surgical progression in an ex vivo specimen. We compute the error between correspondences from the updated model and ground-truth intraoperative CT in the region of anatomical modification. The resulting models show a decrease in error during surgical progression as opposed to increasing when no update is employed. Conclusion: Our findings suggest that preoperative 3D anatomical models can be updated using intraoperative endoscopy video in navigated sinus surgery. Future work will investigate improvements to monocular depth estimation as well as removing the need for external navigation systems. The resulting ability to continuously update the patient model may provide surgeons with a more precise understanding of the current anatomical state and paves the way toward a digital twin paradigm for sinus surgery.

Han Zhang, Lalithkumar Seenivasan, Jose L. Porras et al.

Observing surgical practice has historically relied on fixed vantage points or recollections, leaving the egocentric visual perspectives that guide clinical decisions undocumented. Fixed-camera video can capture surgical workflows at the room-scale, but cannot reconstruct what each team member actually saw. Thus, these videos only provide limited insights into how decisions that affect surgical safety, training, and workflow optimization are made. Here we introduce EgoSurg, the first framework to reconstruct the dynamic, egocentric replays for any operating room (OR) staff directly from wall-mounted fixed-camera video, and thus, without intervention to clinical workflow. EgoSurg couples geometry-driven neural rendering with diffusion-based view enhancement, enabling high-visual fidelity synthesis of arbitrary and egocentric viewpoints at any moment. In evaluation across multi-site surgical cases and controlled studies, EgoSurg reconstructs person-specific visual fields and arbitrary viewpoints with high visual quality and fidelity. By transforming existing OR camera infrastructure into a navigable dynamic 3D record, EgoSurg establishes a new foundation for immersive surgical data science, enabling surgical practice to be visualized, experienced, and analyzed from every angle.