Elena De Momi

Mona Sheikh Zeinoddin, Chiara Lena, Jiongqi Qu et al.

Robotic-assisted surgery (RAS) relies on accurate depth estimation for 3D reconstruction and visualization. While foundation models like Depth Anything Models (DAM) show promise, directly applying them to surgery often yields suboptimal results. Fully fine-tuning on limited surgical data can cause overfitting and catastrophic forgetting, compromising model robustness and generalization. Although Low-Rank Adaptation (LoRA) addresses some adaptation issues, its uniform parameter distribution neglects the inherent feature hierarchy, where earlier layers, learning more general features, require more parameters than later ones. To tackle this issue, we introduce Depth Anything in Robotic Endoscopic Surgery (DARES), a novel approach that employs a new adaptation technique, Vector Low-Rank Adaptation (Vector-LoRA) on the DAM V2 to perform self-supervised monocular depth estimation in RAS scenes. To enhance learning efficiency, we introduce Vector-LoRA by integrating more parameters in earlier layers and gradually decreasing parameters in later layers. We also design a reprojection loss based on the multi-scale SSIM error to enhance depth perception by better tailoring the foundation model to the specific requirements of the surgical environment. The proposed method is validated on the SCARED dataset and demonstrates superior performance over recent state-of-the-art self-supervised monocular depth estimation techniques, achieving an improvement of 13.3% in the absolute relative error metric. The code and pre-trained weights are available at https://github.com/mobarakol/DARES.

Luca Fortini, Mattia Leonori, Juan M. Gandarias et al.

Simulation tools are essential for robotics research, especially for those domains in which safety is crucial, such as Human-Robot Collaboration (HRC). However, it is challenging to simulate human behaviors, and existing robotics simulators do not integrate functional human models. This work presents Open-VICO, an open-source toolkit to integrate virtual human models in Gazebo focusing on vision-based human tracking. In particular, Open-VICO allows to combine in the same simulation environment realistic human kinematic models, multi-camera vision setups, and human-tracking techniques along with numerous robot and sensor models thanks to Gazebo. The possibility to incorporate pre-recorded human skeleton motion with Motion Capture systems broadens the landscape of human performance behavioral analysis within Human-Robot Interaction (HRI) settings. To describe the functionalities and stress the potential of the toolkit four specific examples, chosen among relevant literature challenges in the field, are developed using our simulation utils: i) 3D multi-RGB-D camera calibration in simulation, ii) creation of a synthetic human skeleton tracking dataset based on OpenPose, iii) multi-camera scenario for human skeleton tracking in simulation, and iv) a human-robot interaction example. The key of this work is to create a straightforward pipeline which we hope will motivate research on new vision-based algorithms and methodologies for lightweight human-tracking and flexible human-robot applications.

Francesca Fati, Alberto Rota, Adriana V. Gregory et al.

Adnexal mass evaluation via ultrasound is a challenging clinical task, often hindered by subjective interpretation and significant inter-observer variability. While automated segmentation is a foundational step for quantitative risk assessment, traditional fully supervised convolutional architectures frequently require large amounts of pixel-level annotations and struggle with domain shifts common in medical imaging. In this work, we propose a label-efficient segmentation framework that leverages the robust semantic priors of a pretrained DINOv3 foundational vision transformer backbone. By integrating this backbone with a Dense Prediction Transformer (DPT)-style decoder, our model hierarchically reassembles multi-scale features to combine global semantic representations with fine-grained spatial details. Evaluated on a clinical dataset of 7,777 annotated frames from 112 patients, our method achieves state-of-the-art performance compared to established fully supervised baselines, including U-Net, U-Net++, DeepLabV3, and MAnet. Specifically, we obtain a Dice score of 0.945 and improved boundary adherence, reducing the 95th-percentile Hausdorff Distance by 11.4% relative to the strongest convolutional baseline. Furthermore, we conduct an extensive efficiency analysis demonstrating that our DINOv3-based approach retains significantly higher performance under data starvation regimes, maintaining strong results even when trained on only 25% of the data. These results suggest that leveraging large-scale self-supervised foundations provides a promising and data-efficient solution for medical image segmentation in data-constrained clinical environments. Project Repository: https://github.com/FrancescaFati/MESA

Mauro Orazio Drago, Luca Carlini, Pelinsu Celebi Balyemez et al.

Video Question Answering (VideoQA) in the surgical domain aims to enhance intraoperative understanding by enabling AI models to reason over temporally coherent events rather than isolated frames. Current approaches are limited to static image features, and available datasets often lack temporal annotations, ignoring the dynamics critical for accurate procedural interpretation. We propose SurgViVQA, a surgical VideoQA model that extends visual reasoning from static images to dynamic surgical scenes. It uses a Masked Video--Text Encoder to fuse video and question features, capturing temporal cues such as motion and tool--tissue interactions, which a fine-tuned large language model (LLM) then decodes into coherent answers. To evaluate its performance, we curated REAL-Colon-VQA, a colonoscopic video dataset that includes motion-related questions and diagnostic attributes, as well as out-of-template questions with rephrased or semantically altered formulations to assess model robustness. Experimental validation on REAL-Colon-VQA and the public EndoVis18-VQA dataset shows that SurgViVQA outperforms existing image-based VQA benchmark models, particularly in keyword accuracy, improving over PitVQA by +11\% on REAL-Colon-VQA and +9\% on EndoVis18-VQA. A perturbation study on the questions further confirms improved generalizability and robustness to variations in question phrasing. SurgViVQA and the REAL-Colon-VQA dataset provide a framework for temporally-aware understanding in surgical VideoQA, enabling AI models to interpret dynamic procedural contexts more effectively. Code and dataset available at https://github.com/madratak/SurgViVQA.

Dennis Pierantozzi, Luca Carlini, Mauro Orazio Drago et al.

Safety and reliability are essential for deploying Visual Question Answering (VQA) in surgery, where incorrect or ambiguous responses can harm the patient. Most surgical VQA research focuses on accuracy or linguistic quality while overlooking safety behaviors such as ambiguity awareness, referral to human experts, or triggering a second opinion. Inspired by Automatic Failure Detection (AFD), we study uncertainty estimation as a key enabler of safer decision making. We introduce Question Aligned Semantic Nearest Neighbor Entropy (QA-SNNE), a black box uncertainty estimator that incorporates question semantics into prediction confidence. It measures semantic entropy by comparing generated answers with nearest neighbors in a medical text embedding space, conditioned on the question. We evaluate five models, including domain specific Parameter-Efficient Fine-Tuned (PEFT) models and zero-shot Large Vision-Language Models (LVLMs), on EndoVis18-VQA and PitVQA. PEFT models degrade under mild paraphrasing, while LVLMs are more resilient. Across three LVLMs and two PEFT baselines, QA-SNNE improves AUROC in most in-template settings and enhances hallucination detection. The Area Under the ROC Curve (AUROC) increases by 15-38% for zero-shot models, with gains maintained under out-of-template stress. QA-SNNE offers a practical and interpretable step toward AFD in surgical VQA by linking semantic uncertainty to question context. Combining LVLM backbones with question aligned uncertainty estimation can improve safety and clinician trust. The code and model are available at https://github.com/DennisPierantozzi/QASNNE

Sophia Bano, Alessandro Casella, Francisco Vasconcelos et al.

Fetoscopy laser photocoagulation is a widely adopted procedure for treating Twin-to-Twin Transfusion Syndrome (TTTS). The procedure involves photocoagulation pathological anastomoses to regulate blood exchange among twins. The procedure is particularly challenging due to the limited field of view, poor manoeuvrability of the fetoscope, poor visibility, and variability in illumination. These challenges may lead to increased surgery time and incomplete ablation. Computer-assisted intervention (CAI) can provide surgeons with decision support and context awareness by identifying key structures in the scene and expanding the fetoscopic field of view through video mosaicking. Research in this domain has been hampered by the lack of high-quality data to design, develop and test CAI algorithms. Through the Fetoscopic Placental Vessel Segmentation and Registration (FetReg2021) challenge, which was organized as part of the MICCAI2021 Endoscopic Vision challenge, we released the first largescale multicentre TTTS dataset for the development of generalized and robust semantic segmentation and video mosaicking algorithms. For this challenge, we released a dataset of 2060 images, pixel-annotated for vessels, tool, fetus and background classes, from 18 in-vivo TTTS fetoscopy procedures and 18 short video clips. Seven teams participated in this challenge and their model performance was assessed on an unseen test dataset of 658 pixel-annotated images from 6 fetoscopic procedures and 6 short clips. The challenge provided an opportunity for creating generalized solutions for fetoscopic scene understanding and mosaicking. In this paper, we present the findings of the FetReg2021 challenge alongside reporting a detailed literature review for CAI in TTTS fetoscopy. Through this challenge, its analysis and the release of multi-centre fetoscopic data, we provide a benchmark for future research in this field.

Alessandro Casella, Sophia Bano, Francisco Vasconcelos et al.

In Twin-to-Twin Transfusion Syndrome (TTTS), abnormal vascular anastomoses in the monochorionic placenta can produce uneven blood flow between the two fetuses. In the current practice, TTTS is treated surgically by closing abnormal anastomoses using laser ablation. This surgery is minimally invasive and relies on fetoscopy. Limited field of view makes anastomosis identification a challenging task for the surgeon. To tackle this challenge, we propose a learning-based framework for in-vivo fetoscopy frame registration for field-of-view expansion. The novelties of this framework relies on a learning-based keypoint proposal network and an encoding strategy to filter (i) irrelevant keypoints based on fetoscopic image segmentation and (ii) inconsistent homographies. We validate of our framework on a dataset of 6 intraoperative sequences from 6 TTTS surgeries from 6 different women against the most recent state of the art algorithm, which relies on the segmentation of placenta vessels. The proposed framework achieves higher performance compared to the state of the art, paving the way for robust mosaicking to provide surgeons with context awareness during TTTS surgery.

Jorge F. Lazo, Chun-Feng Lai, Sara Moccia et al.

Navigation inside luminal organs is an arduous task that requires non-intuitive coordination between the movement of the operator's hand and the information obtained from the endoscopic video. The development of tools to automate certain tasks could alleviate the physical and mental load of doctors during interventions, allowing them to focus on diagnosis and decision-making tasks. In this paper, we present a synergic solution for intraluminal navigation consisting of a 3D printed endoscopic soft robot that can move safely inside luminal structures. Visual servoing, based on Convolutional Neural Networks (CNNs) is used to achieve the autonomous navigation task. The CNN is trained with phantoms and in-vivo data to segment the lumen, and a model-less approach is presented to control the movement in constrained environments. The proposed robot is validated in anatomical phantoms in different path configurations. We analyze the movement of the robot using different metrics such as task completion time, smoothness, error in the steady-state, and mean and maximum error. We show that our method is suitable to navigate safely in hollow environments and conditions which are different than the ones the network was originally trained on.

Jorge F. Lazo, Benoit Rosa, Michele Catellani et al.

Objective: Accurate visual classification of bladder tissue during Trans-Urethral Resection of Bladder Tumor (TURBT) procedures is essential to improve early cancer diagnosis and treatment. During TURBT interventions, White Light Imaging (WLI) and Narrow Band Imaging (NBI) techniques are used for lesion detection. Each imaging technique provides diverse visual information that allows clinicians to identify and classify cancerous lesions. Computer vision methods that use both imaging techniques could improve endoscopic diagnosis. We address the challenge of tissue classification when annotations are available only in one domain, in our case WLI, and the endoscopic images correspond to an unpaired dataset, i.e. there is no exact equivalent for every image in both NBI and WLI domains. Method: We propose a semi-surprised Generative Adversarial Network (GAN)-based method composed of three main components: a teacher network trained on the labeled WLI data; a cycle-consistency GAN to perform unpaired image-to-image translation, and a multi-input student network. To ensure the quality of the synthetic images generated by the proposed GAN we perform a detailed quantitative, and qualitative analysis with the help of specialists. Conclusion: The overall average classification accuracy, precision, and recall obtained with the proposed method for tissue classification are 0.90, 0.88, and 0.89 respectively, while the same metrics obtained in the unlabeled domain (NBI) are 0.92, 0.64, and 0.94 respectively. The quality of the generated images is reliable enough to deceive specialists. Significance: This study shows the potential of using semi-supervised GAN-based bladder tissue classification when annotations are limited in multi-domain data. The dataset is available at https://zenodo.org/record/7741476#.ZBQUK7TMJ6k

Ziheng Wang, Andrea Mariani, Arianna Menciassi et al.

Objective technical skill assessment is crucial for effective training of new surgeons in robot-assisted surgery. With advancements in surgical training programs in both physical and virtual environments, it is imperative to develop generalizable methods for automatically assessing skills. In this paper, we propose a novel approach for skill assessment by transferring domain knowledge from labeled kinematic data to unlabeled data. Our approach leverages labeled data from common surgical training tasks such as Suturing, Needle Passing, and Knot Tying to jointly train a model with both labeled and unlabeled data. Pseudo labels are generated for the unlabeled data through an iterative manner that incorporates uncertainty estimation to ensure accurate labeling. We evaluate our method on a virtual reality simulated training task (Ring Transfer) using data from the da Vinci Research Kit (dVRK). The results show that trainees with robotic assistance have significantly higher expert probability compared to these without any assistance, p < 0.05, which aligns with previous studies showing the benefits of robotic assistance in improving training proficiency. Our method offers a significant advantage over other existing works as it does not require manual labeling or prior knowledge of the surgical training task for robot-assisted surgery.

Luca Fortini, Mattia Leonori, Juan M. Gandarias et al.

Tracking 3D human motion in real-time is crucial for numerous applications across many fields. Traditional approaches involve attaching artificial fiducial objects or sensors to the body, limiting their usability and comfort-of-use and consequently narrowing their application fields. Recent advances in Artificial Intelligence (AI) have allowed for markerless solutions. However, most of these methods operate in 2D, while those providing 3D solutions compromise accuracy and real-time performance. To address this challenge and unlock the potential of visual pose estimation methods in real-world scenarios, we propose a markerless framework that combines multi-camera views and 2D AI-based pose estimation methods to track 3D human motion. Our approach integrates a Weighted Least Square (WLS) algorithm that computes 3D human motion from multiple 2D pose estimations provided by an AI-driven method. The method is integrated within the Open-VICO framework allowing simulation and real-world execution. Several experiments have been conducted, which have shown high accuracy and real-time performance, demonstrating the high level of readiness for real-world applications and the potential to revolutionize human motion capture.

Luca Sestini, Benoit Rosa, Elena De Momi et al.

Instance segmentation of surgical instruments is a long-standing research problem, crucial for the development of many applications for computer-assisted surgery. This problem is commonly tackled via fully-supervised training of deep learning models, requiring expensive pixel-level annotations to train. In this work, we develop a framework for instance segmentation not relying on spatial annotations for training. Instead, our solution only requires binary tool masks, obtainable using recent unsupervised approaches, and binary tool presence labels, freely obtainable in robot-assisted surgery. Based on the binary mask information, our solution learns to extract individual tool instances from single frames, and to encode each instance into a compact vector representation, capturing its semantic features. Such representations guide the automatic selection of a tiny number of instances (8 only in our experiments), displayed to a human operator for tool-type labelling. The gathered information is finally used to match each training instance with a binary tool presence label, providing an effective supervision signal to train a tool instance classifier. We validate our framework on the EndoVis 2017 and 2018 segmentation datasets. We provide results using binary masks obtained either by manual annotation or as predictions of an unsupervised binary segmentation model. The latter solution yields an instance segmentation approach completely free from spatial annotations, outperforming several state-of-the-art fully-supervised segmentation approaches.

Francesca Fati, Felipe Coutinho, Marika Reinius et al.

Purpose. High-grade serous ovarian carcinoma (HGSOC) is characterized by pronounced biological and spatial heterogeneity and is frequently diagnosed at an advanced stage. Neoadjuvant chemotherapy (NACT) followed by delayed primary surgery is commonly employed in patients unsuitable for primary cytoreduction. The Chemotherapy Response Score (CRS) is a validated histopathological biomarker of response to NACT, but it is only available postoperatively. In this study, we investigate whether pre-treatment computed tomography (CT) imaging and clinical data can be used to predict CRS as an investigational decision-support adjunct to inform multidisciplinary team (MDT) discussions regarding expected treatment response. Methods. We proposed a 2.5D multimodal deep learning framework that processes lesion-dense omental slices using a pre-trained Vision Transformer encoder and integrates the resulting visual representations with clinical variables through an intermediate fusion module to predict CRS. Results. Our multimodal model, integrating imaging and clinical data, achieved a ROC-AUC of 0.95 alongside 95% accuracy and 80% precision on the internal test cohort (IEO, n=41 patients). On the external test set (OV04, n=70 patients), it achieved a ROC-AUC of 0.68, alongside 67% accuracy and 75% precision. Conclusion. These preliminary results demonstrate the feasibility of transformer-based deep learning for preoperative prediction of CRS in HGSOC using routine clinical data and CT imaging. As an investigational, pre-treatment decision-support tool, this approach may assist MDT discussions by providing early, non-invasive estimates of treatment response.

Alberto Rota, Mert Kiray, Mert Asim Karaoglu et al.

Specular highlights distort appearance, obscure texture, and hinder geometric reasoning in both natural and surgical imagery. We present UnReflectAnything, an RGB-only framework that removes highlights from a single image by predicting a highlight map together with a reflection-free diffuse reconstruction. The model uses a frozen vision transformer encoder to extract multi-scale features, a lightweight head to localize specular regions, and a token-level inpainting module that restores corrupted feature patches before producing the final diffuse image. To overcome the lack of paired supervision, we introduce a Virtual Highlight Synthesis pipeline that renders physically plausible specularities using monocular geometry, Fresnel-aware shading, and randomized lighting which enables training on arbitrary RGB images with correct geometric structure. UnReflectAnything generalizes across natural and surgical domains where non-Lambertian surfaces and non-uniform lighting create severe highlights and it achieves competitive performance with state-of-the-art results on several benchmarks. Project Page: https://alberto-rota.github.io/UnReflectAnything/

Edoardo Lamon, Fabio Fusaro, Elena De Momi et al.

The growing deployment of human-robot collaborative processes in several industrial applications, such as handling, welding, and assembly, unfolds the pursuit of systems which are able to manage large heterogeneous teams and, at the same time, monitor the execution of complex tasks. In this paper, we present a novel architecture for dynamic role allocation and collaborative task planning in a mixed human-robot team of arbitrary size. The architecture capitalizes on a centralized reactive and modular task-agnostic planning method based on Behavior Trees (BTs), in charge of actions scheduling, while the allocation problem is formulated through a Mixed-Integer Linear Program (MILP), that assigns dynamically individual roles or collaborations to the agents of the team. Different metrics used as MILP cost allow the architecture to favor various aspects of the collaboration (e.g. makespan, ergonomics, human preferences). Human preference are identified through a negotiation phase, in which, an human agent can accept/refuse to execute the assigned task.In addition, bilateral communication between humans and the system is achieved through an Augmented Reality (AR) custom user interface that provides intuitive functionalities to assist and coordinate workers in different action phases. The computational complexity of the proposed methodology outperforms literature approaches in industrial sized jobs and teams (problems up to 50 actions and 20 agents in the team with collaborations are solved within 1 s). The different allocated roles, as the cost functions change, highlights the flexibility of the architecture to several production requirements. Finally, the subjective evaluation demonstrating the high usability level and the suitability for the targeted scenario.

Federico Mariano, Elena De Momi, Giovanni Berselli et al.

This paper presents an off-axis, monolithic compliant Remote Center of Motion (RCM) joint for neuroendoscopic manipulation, combining near-isotropic stiffness with minimal parasitic motion. Based on the Tetra II concept, the end-effector is placed outside the tetrahedral flexure to improve line of sight, facilitate sterilization, and allow rapid tool release. Design proceeds in two stages: mobility panels are sized with a compliance-based isotropy objective, then constraining panels are synthesized through finite-element feasibility exploration to trade stiffness isotropy against RCM drift. The joint is modeled with beam elements and validated via detailed finite-element analyses, including fatigue-bounded stress constraints. A PA12 prototype is fabricated by selective laser sintering and characterized on a benchtop: a 2 N radial load is applied at the end-effector while a 6-DOF electromagnetic sensor records pose. The selected configuration produces a stiffness-ellipse principal axis ratio (PAR) of 1.37 and a parasitic-to-useful rotation ratio (PRR) of 0.63%. Under a 4.5° commanded rotation, the predicted RCM drift remains sub-millimetric (0.015-0.172 mm). Fatigue analysis predicts a usable rotational workspace of 12.1°-34.4° depending on direction. Experiments reproduce the simulated directional stiffness trend with typical deviations of 6-30%, demonstrating a compact, fabrication-ready RCM module for constrained surgical access.

Luca Carlini, Chiara Lena, Cesare Hassan et al.

Surgical Video Question Answering (VideoQA) requires accurate temporal grounding while remaining robust to natural variation in how clinicians phrase questions, where linguistic bias can arise. Standard Parameter Efficient Fine Tuning (PEFT) methods adapt pretrained projections without explicitly modeling frame-to-frame interactions within the adaptation pathway, limiting their ability to exploit sparse temporal evidence. We introduce TemporalDoRA, a video-specific PEFT formulation that extends Weight-Decomposed Low-Rank Adaptation by (i) inserting lightweight temporal Multi-Head Attention (MHA) inside the low-rank bottleneck of the vision encoder and (ii) selectively applying weight decomposition only to the trainable low-rank branch rather than the full adapted weight. This design enables temporally-aware updates while preserving a frozen backbone and stable scaling. By mixing information across frames within the adaptation subspace, TemporalDoRA steers updates toward temporally consistent visual cues and improves robustness with minimal parameter overhead. To benchmark this setting, we present REAL-Colon-VQA, a colonoscopy VideoQA dataset with 6,424 clip--question pairs, including paired rephrased Out-of-Template questions to evaluate sensitivity to linguistic variation. TemporalDoRA improves Out-of-Template performance, and ablation studies confirm that temporal mixing inside the low-rank branch is the primary driver of these gains. We also validate on EndoVis18-VQA adapted to short clips and observe consistent improvements on the Out-of-Template split. Code and dataset available at~\href{https://anonymous.4open.science/r/TemporalDoRA-BFC8/}{Anonymous GitHub}.

Alberto Rota, Elena De Momi

Accurate spatial understanding is essential for image-guided surgery, augmented reality integration and context awareness. In minimally invasive procedures, where visual input is the sole intraoperative modality, establishing precise pixel-level correspondences between endoscopic frames is critical for 3D reconstruction, camera tracking, and scene interpretation. However, the surgical domain presents distinct challenges: weak perspective cues, non-Lambertian tissue reflections, and complex, deformable anatomy degrade the performance of conventional computer vision techniques. While Deep Learning models have shown strong performance in natural scenes, their features are not inherently suited for fine-grained matching in surgical images and require targeted adaptation to meet the demands of this domain. This research presents a novel Deep Learning pipeline for establishing feature correspondences in endoscopic image pairs, alongside a self-supervised optimization framework for model training. The proposed methodology leverages a novel-view synthesis pipeline to generate ground-truth inlier correspondences, subsequently utilized for mining triplets within a contrastive learning paradigm. Through this self-supervised approach, we augment the DINOv2 backbone with an additional Transformer layer, specifically optimized to produce embeddings that facilitate direct matching through cosine similarity thresholding. Experimental evaluation demonstrates that our pipeline surpasses state-of-the-art methodologies on the SCARED datasets improved matching precision and lower epipolar error compared to the related work. The proposed framework constitutes a valuable contribution toward enabling more accurate high-level computer vision applications in surgical endoscopy.

Chiara Lena, Davide Milesi, Alessandro Casella et al.

Deep learning has the potential to improve colonoscopy by enabling 3D reconstruction of the colon, providing a comprehensive view of mucosal surfaces and lesions, and facilitating the identification of unexplored areas. However, the development of robust methods is limited by the scarcity of large-scale ground truth data. We propose RealSynCol, a highly realistic synthetic dataset designed to replicate the endoscopic environment. Colon geometries extracted from 10 CT scans were imported into a virtual environment that closely mimics intraoperative conditions and rendered with realistic vascular textures. The resulting dataset comprises 28\,130 frames, paired with ground truth depth maps, optical flow, 3D meshes, and camera trajectories. A benchmark study was conducted to evaluate the available synthetic colon datasets for the tasks of depth and pose estimation. Results demonstrate that the high realism and variability of RealSynCol significantly enhance generalization performance on clinical images, proving it to be a powerful tool for developing deep learning algorithms to support endoscopic diagnosis.

Francesco Pastori, Francesca Fati, Marina Rosanu et al.

Ovarian cancer is the most lethal gynecologic malignancy: around 60% of patients are diagnosed at an advanced stage, with an associated 5-year survival rate of about 30%. Early identification of non-responders to neoadjuvant chemotherapy remains a key unmet need, as it could prevent ineffective therapy and avoid delays in optimal surgical management. This work proposes a non-invasive deep learning framework to predict neoadjuvant chemotherapy response from pre-treatment contrast-enhanced CT by leveraging automatically derived 3D lesion masks. The approach encodes axial slices with a partially fine-tuned pretrained image encoder and aggregates slice-level representations into a volumetric embedding through an attention-based module. Training combines classification loss with supervised contrastive regularization and hard-negative mining to improve separation between ambiguous responders and non-responders. The method was developed on a retrospective single-center cohort from the European Institute of Oncology (Milan, IT), including 280 eligible patients (147 responder, 133 non-responder). On the test cohort, the model achieved a ROC-AUC of 0.73 (95% CI: 0.58-0.86) and an F1-score of 0.70 (95% CI: 0.56-0.82). Overall, these results suggest that the proposed architecture learns clinically relevant predictive patterns and provides a robust foundation for an imaging-based stratification tool.

Yin Lin, Riccardo Barbieri, Domenico Aquino et al.

Glioblastoma is one of the most aggressive and common brain tumors, with a median survival of 10-15 months. Predicting Overall Survival (OS) is critical for personalizing treatment strategies and aligning clinical decisions with patient outcomes. In this study, we propose a novel Artificial Intelligence (AI) approach for OS prediction using Magnetic Resonance Imaging (MRI) images, exploiting Vision Transformers (ViTs) to extract hidden features directly from MRI images, eliminating the need of tumor segmentation. Unlike traditional approaches, our method simplifies the workflow and reduces computational resource requirements. The proposed model was evaluated on the BRATS dataset, reaching an accuracy of 62.5% on the test set, comparable to the top-performing methods. Additionally, it demonstrated balanced performance across precision, recall, and F1 score, overcoming the best model in these metrics. The dataset size limits the generalization of the ViT which typically requires larger datasets compared to convolutional neural networks. This limitation in generalization is observed across all the cited studies. This work highlights the applicability of ViTs for downsampled medical imaging tasks and establishes a foundation for OS prediction models that are computationally efficient and do not rely on segmentation.

Francesco Iodice, Elena De Momi, Arash Ajoudani

The integration of collaborative robots into industrial environments has improved productivity, but has also highlighted significant challenges related to operator safety and ergonomics. This paper proposes an innovative framework that integrates advanced visual perception, continuous ergonomic monitoring, and adaptive Behaviour Tree decision-making to overcome the limitations of traditional methods that typically operate as isolated components. Our approach synthesizes deep learning models, advanced tracking algorithms, and dynamic ergonomic assessments into a modular, scalable, and adaptive system. Experimental validation demonstrates the framework's superiority over existing solutions across multiple dimensions: the visual perception module outperformed previous detection models with 72.4% mAP@50:95; the system achieved high accuracy in recognizing operator intentions (92.5%); it promptly classified ergonomic risks with minimal latency (0.57 seconds); and it dynamically managed robotic interventions with exceptionally responsive decision-making capabilities (0.07 seconds), representing a 56% improvement over benchmark systems. This comprehensive solution provides a robust platform for enhancing human-robot collaboration in industrial environments by prioritizing ergonomic safety, operational efficiency, and real-time adaptability.

Luca Sestini, Benoit Rosa, Elena De Momi et al.

Automatic surgical instrument segmentation of endoscopic images is a crucial building block of many computer-assistance applications for minimally invasive surgery. So far, state-of-the-art approaches completely rely on the availability of a ground-truth supervision signal, obtained via manual annotation, thus expensive to collect at large scale. In this paper, we present FUN-SIS, a Fully-UNsupervised approach for binary Surgical Instrument Segmentation. FUN-SIS trains a per-frame segmentation model on completely unlabelled endoscopic videos, by solely relying on implicit motion information and instrument shape-priors. We define shape-priors as realistic segmentation masks of the instruments, not necessarily coming from the same dataset/domain as the videos. The shape-priors can be collected in various and convenient ways, such as recycling existing annotations from other datasets. We leverage them as part of a novel generative-adversarial approach, allowing to perform unsupervised instrument segmentation of optical-flow images during training. We then use the obtained instrument masks as pseudo-labels in order to train a per-frame segmentation model; to this aim, we develop a learning-from-noisy-labels architecture, designed to extract a clean supervision signal from these pseudo-labels, leveraging their peculiar noise properties. We validate the proposed contributions on three surgical datasets, including the MICCAI 2017 EndoVis Robotic Instrument Segmentation Challenge dataset. The obtained fully-unsupervised results for surgical instrument segmentation are almost on par with the ones of fully-supervised state-of-the-art approaches. This suggests the tremendous potential of the proposed method to leverage the great amount of unlabelled data produced in the context of minimally invasive surgery.

Marta Lagomarsino, Marta Lorenzini, Elena De Momi et al.

The ongoing trend towards Industry 4.0 has revolutionised ordinary workplaces, profoundly changing the role played by humans in the production chain. Research on ergonomics in industrial settings mainly focuses on reducing the operator's physical fatigue and discomfort to improve throughput and avoid safety hazards. However, as the production complexity increases, the cognitive resources demand and mental workload could compromise the operator's performance and the efficiency of the shop floor workplace. State-of-the-art methods in cognitive science work offline and/or involve bulky equipment hardly deployable in industrial settings. This paper presents a novel method for online assessment of cognitive load in manufacturing, primarily assembly, by detecting patterns in human motion directly from the input images of a stereo camera. Head pose estimation and skeleton tracking are exploited to investigate the workers' attention and assess hyperactivity and unforeseen movements. Pilot experiments suggest that our factor assessment tool provides significant insights into workers' mental workload, even confirmed by correlations with physiological and performance measurements. According to data gathered in this study, a vision-based cognitive load assessment has the potential to be integrated into the development of mechatronic systems for improving cognitive ergonomics in manufacturing.

Alice Segato, Chiara Di Vece, Sara Zucchelli et al.

Many tasks in robot-assisted surgery require planning and controlling manipulators' motions that interact with highly deformable objects. This study proposes a realistic, time-bounded simulator based on Position-based Dynamics (PBD) simulation that mocks brain deformations due to catheter insertion for pre-operative path planning and intra-operative guidance in keyhole surgical procedures. It maximizes the probability of success by accounting for uncertainty in deformation models, noisy sensing, and unpredictable actuation. The PBD deformation parameters were initialized on a parallelepiped-shaped simulated phantom to obtain a reasonable starting guess for the brain white matter. They were calibrated by comparing the obtained displacements with deformation data for catheter insertion in a composite hydrogel phantom. Knowing the gray matter brain structures' different behaviors, the parameters were fine-tuned to obtain a generalized human brain model. The brain structures' average displacement was compared with values in the literature. The simulator's numerical model uses a novel approach with respect to the literature, and it has proved to be a close match with real brain deformations through validation using recorded deformation data of in-vivo animal trials with a mean mismatch of 4.73$\pm$2.15%. The stability, accuracy, and real-time performance make this model suitable for creating a dynamic environment for KN path planning, pre-operative path planning, and intra-operative guidance.

Sophia Bano, Alessandro Casella, Francisco Vasconcelos et al.

Fetoscopy laser photocoagulation is a widely used procedure for the treatment of Twin-to-Twin Transfusion Syndrome (TTTS), that occur in mono-chorionic multiple pregnancies due to placental vascular anastomoses. This procedure is particularly challenging due to limited field of view, poor manoeuvrability of the fetoscope, poor visibility due to fluid turbidity, variability in light source, and unusual position of the placenta. This may lead to increased procedural time and incomplete ablation, resulting in persistent TTTS. Computer-assisted intervention may help overcome these challenges by expanding the fetoscopic field of view through video mosaicking and providing better visualization of the vessel network. However, the research and development in this domain remain limited due to unavailability of high-quality data to encode the intra- and inter-procedure variability. Through the \textit{Fetoscopic Placental Vessel Segmentation and Registration (FetReg)} challenge, we present a large-scale multi-centre dataset for the development of generalized and robust semantic segmentation and video mosaicking algorithms for the fetal environment with a focus on creating drift-free mosaics from long duration fetoscopy videos. In this paper, we provide an overview of the FetReg dataset, challenge tasks, evaluation metrics and baseline methods for both segmentation and registration. Baseline methods results on the FetReg dataset shows that our dataset poses interesting challenges, offering large opportunity for the creation of novel methods and models through a community effort initiative guided by the FetReg challenge.

Fabio Fusaro, Edoardo Lamon, Elena De Momi et al.

This paper proposes a novel integrated dynamic method based on Behavior Trees for planning and allocating tasks in mixed human robot teams, suitable for manufacturing environments. The Behavior Tree formulation allows encoding a single job as a compound of different tasks with temporal and logic constraints. In this way, instead of the well-studied offline centralized optimization problem, the role allocation problem is solved with multiple simplified online optimization sub-problem, without complex and cross-schedule task dependencies. These sub-problems are defined as Mixed-Integer Linear Programs, that, according to the worker-actions related costs and the workers' availability, allocate the yet-to-execute tasks among the available workers. To characterize the behavior of the developed method, we opted to perform different simulation experiments in which the results of the action-worker allocation and computational complexity are evaluated. The obtained results, due to the nature of the algorithm and to the possibility of simulating the agents' behavior, should describe well also how the algorithm performs in real experiments.

Soheil Gholami, Marta Lorenzini, Elena De Momi et al.

Human factors and ergonomics are the essential constituents of teleoperation interfaces, which can significantly affect the human operator's performance. Thus, a quantitative evaluation of these elements and the ability to establish reliable comparison bases for different teleoperation interfaces are the keys to select the most suitable one for a particular application. However, most of the works on teleoperation have so far focused on the stability analysis and the transparency improvement of these systems, and do not cover the important usability aspects. In this work, we propose a foundation to build a general framework for the analysis of human factors and ergonomics in employing diverse teleoperation interfaces. The proposed framework will go beyond the traditional subjective analyses of usability by complementing it with online measurements of the human body configurations. As a result, multiple quantitative metrics such as joints' usage, range of motion comfort, center of mass divergence, and posture comfort are introduced. To demonstrate the potential of the proposed framework, two different teleoperation interfaces are considered, and real-world experiments with eleven participants performing a simulated industrial remote pick-and-place task are conducted. The quantitative results of this analysis are provided, and compared with subjective questionnaires, illustrating the effectiveness of the proposed framework.

Jorge F. Lazo, Sara Moccia, Aldo Marzullo et al.

Ureteroscopy and cystoscopy are the gold standard methods to identify and treat tumors along the urinary tract. It has been reported that during a normal procedure a rate of 10-20 % of the lesions could be missed. In this work we study the implementation of 3 different Convolutional Neural Networks (CNNs), using a 2-steps training strategy, to classify images from the urinary tract with and without lesions. A total of 6,101 images from ureteroscopy and cystoscopy procedures were collected. The CNNs were trained and tested using transfer learning in a two-steps fashion on 3 datasets. The datasets used were: 1) only ureteroscopy images, 2) only cystoscopy images and 3) the combination of both of them. For cystoscopy data, VGG performed better obtaining an Area Under the ROC Curve (AUC) value of 0.846. In the cases of ureteroscopy and the combination of both datasets, ResNet50 achieved the best results with AUC values of 0.987 and 0.940. The use of a training dataset that comprehends both domains results in general better performances, but performing a second stage of transfer learning achieves comparable ones. There is no single model which performs better in all scenarios, but ResNet50 is the network that achieves the best performances in most of them. The obtained results open the opportunity for further investigation with a view for improving lesion detection in endoscopic images of the urinary system.

Jorge F. Lazo, Aldo Marzullo, Sara Moccia et al.

Purpose: Ureteroscopy is an efficient endoscopic minimally invasive technique for the diagnosis and treatment of upper tract urothelial carcinoma (UTUC). During ureteroscopy, the automatic segmentation of the hollow lumen is of primary importance, since it indicates the path that the endoscope should follow. In order to obtain an accurate segmentation of the hollow lumen, this paper presents an automatic method based on Convolutional Neural Networks (CNNs). Methods: The proposed method is based on an ensemble of 4 parallel CNNs to simultaneously process single and multi-frame information. Of these, two architectures are taken as core-models, namely U-Net based in residual blocks($m_1$) and Mask-RCNN($m_2$), which are fed with single still-frames $I(t)$. The other two models ($M_1$, $M_2$) are modifications of the former ones consisting on the addition of a stage which makes use of 3D Convolutions to process temporal information. $M_1$, $M_2$ are fed with triplets of frames ($I(t-1)$, $I(t)$, $I(t+1)$) to produce the segmentation for $I(t)$. Results: The proposed method was evaluated using a custom dataset of 11 videos (2,673 frames) which were collected and manually annotated from 6 patients. We obtain a Dice similarity coefficient of 0.80, outperforming previous state-of-the-art methods. Conclusion: The obtained results show that spatial-temporal information can be effectively exploited by the ensemble model to improve hollow lumen segmentation in ureteroscopic images. The method is effective also in presence of poor visibility, occasional bleeding, or specular reflections.

Pedro Henrique Suruagy Perrusi, Anna Cazzaniga, Paul Baksic et al.

Control strategies for robotic needle steering in soft tissues must account for complex interactions between the needle and the tissue to achieve accurate needle tip positioning. Recent findings show faster robotic command rate can improve the control stability in realistic scenarios. This study proposes the use of Extreme Learning Machines to provide fast commands for robotic needle steering. A synthetic dataset based on the inverse finite element simulation control framework is used to train the model. Results show the model is capable to infer commands 66% faster than the inverse simulation and reaches acceptable precision even on previously unseen trajectories.

Luca Sestini, Benoit Rosa, Elena De Momi et al.

3-D pose estimation of instruments is a crucial step towards automatic scene understanding in robotic minimally invasive surgery. Although robotic systems can potentially directly provide joint values, this information is not commonly exploited inside the operating room, due to its possible unreliability, limited access and the time-consuming calibration required, especially for continuum robots. For this reason, standard approaches for 3-D pose estimation involve the use of external tracking systems. Recently, image-based methods have emerged as promising, non-invasive alternatives. While many image-based approaches in the literature have shown accurate results, they generally require either a complex iterative optimization for each processed image, making them unsuitable for real-time applications, or a large number of manually-annotated images for efficient learning. In this paper we propose a self-supervised image-based method, exploiting, at training time only, the imprecise kinematic information provided by the robot. In order to avoid introducing time-consuming manual annotations, the problem is formulated as an auto-encoder, smartly bottlenecked by the presence of a physical model of the robotic instruments and surgical camera, forcing a separation between image background and kinematic content. Validation of the method was performed on semi-synthetic, phantom and in-vivo datasets, obtained using a flexible robotized endoscope, showing promising results for real-time image-based 3-D pose estimation of surgical instruments.

Jorge F. Lazo, Aldo Marzullo, Sara Moccia et al.

Ureteroscopy is becoming the first surgical treatment option for the majority of urinary affections. This procedure is performed using an endoscope which provides the surgeon with the visual information necessary to navigate inside the urinary tract. Having in mind the development of surgical assistance systems, that could enhance the performance of surgeon, the task of lumen segmentation is a fundamental part since this is the visual reference which marks the path that the endoscope should follow. This is something that has not been analyzed in ureteroscopy data before. However, this task presents several challenges given the image quality and the conditions itself of ureteroscopy procedures. In this paper, we study the implementation of a Deep Neural Network which exploits the advantage of residual units in an architecture based on U-Net. For the training of these networks, we analyze the use of two different color spaces: gray-scale and RGB data images. We found that training on gray-scale images gives the best results obtaining mean values of Dice Score, Precision, and Recall of 0.73, 0.58, and 0.92 respectively. The results obtained shows that the use of residual U-Net could be a suitable model for further development for a computer-aided system for navigation and guidance through the urinary system.

Jorge F. Lazo, Sara Moccia, Emanuele Frontoni et al.

Breast cancer is one of the deadliest cancer worldwide. Timely detection could reduce mortality rates. In the clinical routine, classifying benign and malignant tumors from ultrasound (US) imaging is a crucial but challenging task. An automated method, which can deal with the variability of data is therefore needed. In this paper, we compared different Convolutional Neural Networks (CNNs) and transfer learning methods for the task of automated breast tumor classification. The architectures investigated in this study were VGG-16 and Inception V3. Two different training strategies were investigated: the first one was using pretrained models as feature extractors and the second one was to fine-tune the pre-trained models. A total of 947 images were used, 587 corresponded to US images of benign tumors and 360 with malignant tumors. 678 images were used for the training and validation process, while 269 images were used for testing the models. Accuracy and Area Under the receiver operating characteristic Curve (AUC) were used as performance metrics. The best performance was obtained by fine tuning VGG-16, with an accuracy of 0.919 and an AUC of 0.934. The obtained results open the opportunity to further investigation with a view of improving cancer detection.

Beatrice van Amsterdam, Hirenkumar Nakawala, Elena De Momi et al.

Kinematic trajectories recorded from surgical robots contain information about surgical gestures and potentially encode cues about surgeon's skill levels. Automatic segmentation of these trajectories into meaningful action units could help to develop new metrics for surgical skill assessment as well as to simplify surgical automation. State-of-the-art methods for action recognition relied on manual labelling of large datasets, which is time consuming and error prone. Unsupervised methods have been developed to overcome these limitations. However, they often rely on tedious parameter tuning and perform less well than supervised approaches, especially on data with high variability such as surgical trajectories. Hence, the potential of weak supervision could be to improve unsupervised learning while avoiding manual annotation of large datasets. In this paper, we used at a minimum one expert demonstration and its ground truth annotations to generate an appropriate initialization for a GMM-based algorithm for gesture recognition. We showed on real surgical demonstrations that the latter significantly outperforms standard task-agnostic initialization methods. We also demonstrated how to improve the recognition accuracy further by redefining the actions and optimising the inputs.

Claudia D'Ettorre, George Dwyer, Xiaofei Du et al.

Robot-assisted laparoscopic prostatectomy (RALP) is a treatment for prostate cancer that involves complete or nerve sparing removal prostate tissue that contains cancer. After removal the bladder neck is successively sutured directly with the urethra. The procedure is called urethrovesical anastomosis and is one of the most dexterity demanding tasks during RALP. Two suturing instruments and a pair of needles are used in combination to perform a running stitch during urethrovesical anastomosis. While robotic instruments provide enhanced dexterity to perform the anastomosis, it is still highly challenging and difficult to learn. In this paper, we presents a vision-guided needle grasping method for automatically grasping the needle that has been inserted into the patient prior to anastomosis. We aim to automatically grasp the suturing needle in a position that avoids hand-offs and immediately enables the start of suturing. The full grasping process can be broken down into: a needle detection algorithm; an approach phase where the surgical tool moves closer to the needle based on visual feedback; and a grasping phase through path planning based on observed surgical practice. Our experimental results show examples of successful autonomous grasping that has the potential to simplify and decrease the operational time in RALP by assisting a small component of urethrovesical anastomosis.