Florian Schiffers

Amil Dravid, Florian Schiffers, Boqing Gong et al.

Despite the surge of deep learning in the past decade, some users are skeptical to deploy these models in practice due to their black-box nature. Specifically, in the medical space where there are severe potential repercussions, we need to develop methods to gain confidence in the models' decisions. To this end, we propose a novel medical imaging generative adversarial framework, medXGAN (medical eXplanation GAN), to visually explain what a medical classifier focuses on in its binary predictions. By encoding domain knowledge of medical images, we are able to disentangle anatomical structure and pathology, leading to fine-grained visualization through latent interpolation. Furthermore, we optimize the latent space such that interpolation explains how the features contribute to the classifier's output. Our method outperforms baselines such as Gradient-Weighted Class Activation Mapping (Grad-CAM) and Integrated Gradients in localization and explanatory ability. Additionally, a combination of the medXGAN with Integrated Gradients can yield explanations more robust to noise. The code is available at: https://avdravid.github.io/medXGAN_page/.

Tobia Poppi, Silvia Cappelletti, Sara Sarto et al.

Recent progress in generative modeling has made safety control a central challenge, yet existing approaches remain largely model-specific, requiring retraining or tailored interventions for each new architecture. In this work, we ask whether safety can be represented as a portable latent direction, learned once and reused across heterogeneous generators. We introduce the first framework for cross-model safety steering, in which a safety direction is estimated in a source LLM from paired safe-unsafe prompts, transported to a target generator through a lightweight alignment fitted on benign data alone, and applied at inference time. Crucially, our pipeline never accesses unsafe data on the target side, isolating whether safety can be transferred through shared representation geometry. Beyond a single global direction, we also identify a multi-vector extension that captures category-specific safety behaviors, enabling more selective control. We evaluate our approach in text-to-image and text-to-video generation across diverse source-target model pairs. Across models, transferred safety directions achieve ASR reduction and CLIP-Score/FID trade-offs comparable to directions learned natively on the target model using unsafe data, while requiring no target-side unsafe data. This indicates that safety improvements do not come at the expense of generation quality. Our results point to a modular view of safety: safety-relevant behavior is not purely model-local, but can be controlled through latent directions that persist across models. This suggests a new path toward lightweight, reusable safety mechanisms that do not require target-side unsafe data.

Florian Schiffers, Praneeth Chakravarthula, Nathan Matsuda et al.

The Visual Turing Test is the ultimate goal to evaluate the realism of holographic displays. Previous studies have focused on addressing challenges such as limited étendue and image quality over a large focal volume, but they have not investigated the effect of pupil sampling on the viewing experience in full 3D holograms. In this work, we tackle this problem with a novel hologram generation algorithm motivated by matching the projection operators of incoherent Light Field and coherent Wigner Function light transport. To this end, we supervise hologram computation using synthesized photographs, which are rendered on-the-fly using Light Field refocusing from stochastically sampled pupil states during optimization. The proposed method produces holograms with correct parallax and focus cues, which are important for passing the Visual Turing Test. We validate that our approach compares favorably to state-of-the-art CGH algorithms that use Light Field and Focal Stack supervision. Our experiments demonstrate that our algorithm significantly improves the realism of the viewing experience for a variety of different pupil states.

Tobia Poppi, Burak Uzkent, Amanmeet Garg et al.

Video-language models (VLMs) achieve strong multimodal understanding but remain prone to hallucinations, especially when reasoning about actions and temporal order. Existing mitigation strategies, such as textual filtering or random video perturbations, often fail to address the root cause: over-reliance on language priors rather than fine-grained visual dynamics. We propose a scalable framework for counterfactual video generation that synthesizes videos differing only in actions or temporal structure while preserving scene context. Our pipeline combines multimodal LLMs for action proposal and editing guidance with diffusion-based image and video models to generate semantic hard negatives at scale. Using this framework, we build CounterVid, a synthetic dataset of ~26k preference pairs targeting action recognition and temporal reasoning. We further introduce MixDPO, a unified Direct Preference Optimization approach that jointly leverages textual and visual preferences. Fine-tuning Qwen2.5-VL with MixDPO yields consistent improvements, notably in temporal ordering, and transfers effectively to standard video hallucination benchmarks. Code and models will be made publicly available.

Hui Lin, Florian Schiffers, Santiago López-Tapia et al.

Unsupervised domain adaptation (UDA) is essential for medical image segmentation, especially in cross-modality data scenarios. UDA aims to transfer knowledge from a labeled source domain to an unlabeled target domain, thereby reducing the dependency on extensive manual annotations. This paper presents DRL-STNet, a novel framework for cross-modality medical image segmentation that leverages generative adversarial networks (GANs), disentangled representation learning (DRL), and self-training (ST). Our method leverages DRL within a GAN to translate images from the source to the target modality. Then, the segmentation model is initially trained with these translated images and corresponding source labels and then fine-tuned iteratively using a combination of synthetic and real images with pseudo-labels and real labels. The proposed framework exhibits superior performance in abdominal organ segmentation on the FLARE challenge dataset, surpassing state-of-the-art methods by 11.4% in the Dice similarity coefficient and by 13.1% in the Normalized Surface Dice metric, achieving scores of 74.21% and 80.69%, respectively. The average running time is 41 seconds, and the area under the GPU memory-time curve is 11,292 MB. These results indicate the potential of DRL-STNet for enhancing cross-modality medical image segmentation tasks.

Florian Schiffers, Grace Kuo, Nathan Matsuda et al.

Holographic displays hold the promise of providing authentic depth cues, resulting in enhanced immersive visual experiences for near-eye applications. However, current holographic displays are hindered by speckle noise, which limits accurate reproduction of color and texture in displayed images. We present HoloChrome, a polychromatic holographic display framework designed to mitigate these limitations. HoloChrome utilizes an ultrafast, wavelength-adjustable laser and a dual-Spatial Light Modulator (SLM) architecture, enabling the multiplexing of a large set of discrete wavelengths across the visible spectrum. By leveraging spatial separation in our dual-SLM setup, we independently manipulate speckle patterns across multiple wavelengths. This novel approach effectively reduces speckle noise through incoherent averaging achieved by wavelength multiplexing. Our method is complementary to existing speckle reduction techniques, offering a new pathway to address this challenge. Furthermore, the use of polychromatic illumination broadens the achievable color gamut compared to traditional three-color primary holographic displays. Our simulations and tabletop experiments validate that HoloChrome significantly reduces speckle noise and expands the color gamut. These advancements enhance the performance of holographic near-eye displays, moving us closer to practical, immersive next-generation visual experiences.

Xinjie Li, Zhimin Chen, Rui Zhao et al.

Recent unified models for joint understanding and generation have significantly advanced visual generation capabilities. However, their focus on conventional tasks like text-to-video generation has left the temporal reasoning potential of unified models largely underexplored. To address this gap, we introduce Next Scene Prediction (NSP), a new task that pushes unified video models toward temporal and causal reasoning. Unlike text-to-video generation, NSP requires predicting plausible futures from preceding context, demanding deeper understanding and reasoning. To tackle this task, we propose a unified framework combining Qwen-VL for comprehension and LTX for synthesis, bridged by a latent query embedding and a connector module. This model is trained in three stages on our newly curated, large-scale NSP dataset: text-to-video pre-training, supervised fine-tuning, and reinforcement learning (via GRPO) with our proposed causal consistency reward. Experiments demonstrate our model achieves state-of-the-art performance on our benchmark, advancing the capability of generalist multimodal systems to anticipate what happens next.

Neda Tavakoli, Amir Ali Rahsepar, Brandon C. Benefield et al.

Background: Late Gadolinium Enhancement (LGE) imaging is the gold standard for assessing myocardial fibrosis and scarring, with left ventricular (LV) LGE extent predicting major adverse cardiac events (MACE). Despite its importance, routine LGE-based LV scar quantification is hindered by labor-intensive manual segmentation and inter-observer variability. Methods: We propose ScarNet, a hybrid model combining a transformer-based encoder from the Medical Segment Anything Model (MedSAM) with a convolution-based U-Net decoder, enhanced by tailored attention blocks. ScarNet was trained on 552 ischemic cardiomyopathy patients with expert segmentations of myocardial and scar boundaries and tested on 184 separate patients. Results: ScarNet achieved robust scar segmentation in 184 test patients, yielding a median Dice score of 0.912 (IQR: 0.863--0.944), significantly outperforming MedSAM (median Dice = 0.046, IQR: 0.043--0.047) and nnU-Net (median Dice = 0.638, IQR: 0.604--0.661). ScarNet demonstrated lower bias (-0.63%) and coefficient of variation (4.3%) compared to MedSAM (bias: -13.31%, CoV: 130.3%) and nnU-Net (bias: -2.46%, CoV: 20.3%). In Monte Carlo simulations with noise perturbations, ScarNet achieved significantly higher scar Dice (0.892 \pm 0.053, CoV = 5.9%) than MedSAM (0.048 \pm 0.112, CoV = 233.3%) and nnU-Net (0.615 \pm 0.537, CoV = 28.7%). Conclusion: ScarNet outperformed MedSAM and nnU-Net in accurately segmenting myocardial and scar boundaries in LGE images. The model exhibited robust performance across diverse image qualities and scar patterns.

Amil Dravid, Florian Schiffers, Yunan Wu et al.

Generative Adversarial Networks (GANs) have shown promise in augmenting datasets and boosting convolutional neural networks' (CNN) performance on image classification tasks. But they introduce more hyperparameters to tune as well as the need for additional time and computational power to train supplementary to the CNN. In this work, we examine the potential for Auxiliary-Classifier GANs (AC-GANs) as a 'one-stop-shop' architecture for image classification, particularly in low data regimes. Additionally, we explore modifications to the typical AC-GAN framework, changing the generator's latent space sampling scheme and employing a Wasserstein loss with gradient penalty to stabilize the simultaneous training of image synthesis and classification. Through experiments on images of varying resolutions and complexity, we demonstrate that AC-GANs show promise in image classification, achieving competitive performance with standard CNNs. These methods can be employed as an 'all-in-one' framework with particular utility in the absence of large amounts of training data.

Ethan Tseng, Grace Kuo, Seung-Hwan Baek et al.

Holographic displays can generate light fields by dynamically modulating the wavefront of a coherent beam of light using a spatial light modulator, promising rich virtual and augmented reality applications. However, the limited spatial resolution of existing dynamic spatial light modulators imposes a tight bound on the diffraction angle. As a result, modern holographic displays possess low étendue, which is the product of the display area and the maximum solid angle of diffracted light. The low étendue forces a sacrifice of either the field-of-view (FOV) or the display size. In this work, we lift this limitation by presenting neural étendue expanders. This new breed of optical elements, which is learned from a natural image dataset, enables higher diffraction angles for ultra-wide FOV while maintaining both a compact form factor and the fidelity of displayed contents to human viewers. With neural étendue expanders, we experimentally achieve 64$\times$ étendue expansion of natural images in full color, expanding the FOV by an order of magnitude horizontally and vertically, with high-fidelity reconstruction quality (measured in PSNR) over 29 dB on retinal-resolution images.

Merlin A. Nau, Florian Schiffers, Yunhao Li et al.

The utilization of computational photography becomes increasingly essential in the medical field. Today, imaging techniques for dermatology range from two-dimensional (2D) color imagery with a mobile device to professional clinical imaging systems measuring additional detailed three-dimensional (3D) data. The latter are commonly expensive and not accessible to a broad audience. In this work, we propose a novel system and software framework that relies only on low-cost (and even mobile) commodity devices present in every household to measure detailed 3D information of the human skin with a 3D-gradient-illumination-based method. We believe that our system has great potential for early-stage diagnosis and monitoring of skin diseases, especially in vastly populated or underdeveloped areas.

Haoyu Wei, Florian Schiffers, Tobias Würfl et al.

Computed tomography is widely used to examine internal structures in a non-destructive manner. To obtain high-quality reconstructions, one typically has to acquire a densely sampled trajectory to avoid angular undersampling. However, many scenarios require a sparse-view measurement leading to streak-artifacts if unaccounted for. Current methods do not make full use of the domain-specific information, and hence fail to provide reliable reconstructions for highly undersampled data. We present a novel framework for sparse-view tomography by decoupling the reconstruction into two steps: First, we overcome its ill-posedness using a super-resolution network, SIN, trained on the sparse projections. The intermediate result allows for a closed-form tomographic reconstruction with preserved details and highly reduced streak-artifacts. Second, a refinement network, PRN, trained on the reconstructions reduces any remaining artifacts. We further propose a light-weight variant of the perceptual-loss that enhances domain-specific information, boosting restoration accuracy. Our experiments demonstrate an improvement over current solutions by 4 dB.

Florian Schiffers, Thomas Bochynek, Andre Aichert et al.

The use of computed tomography (CT) imaging has become of increasing interest to academic areas outside of the field of medical imaging and industrial inspection, e.g., to biology and cultural heritage research. The pecularities of these fields, however, sometimes require that objects need to be imaged on-site, e.g., in field-work conditions or in museum collections. Under these circumstances, it is often not possible to use a commercial device and a custom solution is the only viable option. In order to achieve high image quality under adverse conditions, reliable calibration and trajectory reproduction are usually key requirements for any custom CT scanning system. Here, we introduce the construction of a low-cost disassemblable CT scanner that allows calibration even when trajectory reproduction is not possible due to the limitations imposed by the project conditions. Using 3D-printed in-image calibration phantoms, we compute a projection matrix directly from each captured X-ray projection. We describe our method in detail and show successful tomographic reconstructions of several specimen as proof of concept.

Florian Willomitzer, Chia-Kai Yeh, Vikas Gupta et al.

We introduce a system and methods for the three-dimensional measurement of extended specular surfaces with high surface normal variations. Our system consists only of a mobile hand held device and exploits screen and front camera for Deflectometry-based surface measurements. We demonstrate high quality measurements without the need for an offline calibration procedure. In addition, we develop a multi-view technique to compensate for the small screen of a mobile device so that large surfaces can be densely reconstructed in their entirety. This work is a first step towards developing a self-calibrating Deflectometry procedure capable of taking 3D surface measurements of specular objects in the wild and accessible to users with little to no technical imaging experience.

Gerd Häusler, Aleksandra Milczarek, Markus Schreiter et al.

This document contains the results obtained by the Innovation Management Seminar in winter term 2017. In total 11 ideas have been developed by the team. In the document all 11 ideas show improvements for future applications in ophthalmology. The 11 ideas are AR/VR Glasses with Medical Applications, Augmented Reality Eye Surgery, Game Diagnosis, Intelligent Adapting Glasses, MD Facebook, Medical Crowd Segmentation, Personalized 3D Model of the Human Eye, Photoacoustic Contact Lens, Power Supply Smart Contact Lens, VR-Cornea and Head Mount for Fundus Imaging