Siavash Bigdeli

Simon Narduzzi, Siavash A. Bigdeli, Shih-Chii Liu et al.

Reducing energy consumption is a critical point for neural network models running on edge devices. In this regard, reducing the number of multiply-accumulate (MAC) operations of Deep Neural Networks (DNNs) running on edge hardware accelerators will reduce the energy consumption during inference. Spiking Neural Networks (SNNs) are an example of bio-inspired techniques that can further save energy by using binary activations, and avoid consuming energy when not spiking. The networks can be configured for equivalent accuracy on a task through DNN-to-SNN conversion frameworks but their conversion is based on rate coding therefore the synaptic operations can be high. In this work, we look into different techniques to enforce sparsity on the neural network activation maps and compare the effect of different training regularizers on the efficiency of the optimized DNNs and SNNs.

Nina Weng, Siavash Bigdeli, Eike Petersen et al.

While many studies have assessed the fairness of AI algorithms in the medical field, the causes of differences in prediction performance are often unknown. This lack of knowledge about the causes of bias hampers the efficacy of bias mitigation, as evidenced by the fact that simple dataset balancing still often performs best in reducing performance gaps but is unable to resolve all performance differences. In this work, we investigate the causes of gender bias in machine learning-based chest X-ray diagnosis. In particular, we explore the hypothesis that breast tissue leads to underexposure of the lungs and causes lower model performance. Methodologically, we propose a new sampling method which addresses the highly skewed distribution of recordings per patient in two widely used public datasets, while at the same time reducing the impact of label errors. Our comprehensive analysis of gender differences across diseases, datasets, and gender representations in the training set shows that dataset imbalance is not the sole cause of performance differences. Moreover, relative group performance differs strongly between datasets, indicating important dataset-specific factors influencing male/female group performance. Finally, we investigate the effect of breast tissue more specifically, by cropping out the breasts from recordings, finding that this does not resolve the observed performance gaps. In conclusion, our results indicate that dataset-specific factors, not fundamental physiological differences, are the main drivers of male--female performance gaps in chest X-ray analyses on widely used NIH and CheXpert Dataset.

Yeganeh Ghamary, Victoria Wu, Hooman Vaseli et al.

Ejection fraction (EF) is a crucial metric for assessing cardiac function and diagnosing conditions such as heart failure. Traditionally, EF estimation requires manual tracing and domain expertise, making the process time-consuming and subject to interobserver variability. Most current deep learning methods for EF prediction are black-box models with limited transparency, which reduces clinical trust. Some post-hoc explainability methods have been proposed to interpret the decision-making process after the prediction is made. However, these explanations do not guide the model's internal reasoning and therefore offer limited reliability in clinical applications. To address this, we introduce ProtoEFNet, a novel video-based prototype learning model for continuous EF regression. The model learns dynamic spatiotemporal prototypes that capture clinically meaningful cardiac motion patterns. Additionally, the proposed Prototype Angular Separation (PAS) loss enforces discriminative representations across the continuous EF spectrum. Our experiments on the EchonetDynamic dataset show that ProtoEFNet can achieve accuracy on par with its non-interpretable counterpart while providing clinically relevant insight. The ablation study shows that the proposed loss boosts performance with a 2% increase in F1 score from 77.67$\pm$2.68 to 79.64$\pm$2.10. Our source code is available at: https://github.com/DeepRCL/ProtoEF

Pawel Tomasz Pieta, Anders Bjorholm Dahl, Jeppe Revall Frisvad et al.

The structure tensor method is often used for 2D and 3D analysis of imaged structures, but its results are in many cases very dependent on the user's choice of method parameters. We simplify this parameter choice in first order structure tensor scale-space by directly connecting the width of the derivative filter to the size of image features. By introducing a ring-filter step, we substitute the Gaussian integration/smoothing with a method that more accurately shifts the derivative filter response from feature edges to their center. We further demonstrate how extracted structural measures can be used to correct known inaccuracies in the scale map, resulting in a reliable representation of the feature sizes both in 2D and 3D. Compared to the traditional first order structure tensor, or previous structure tensor scale-space approaches, our solution is much more accurate and can serve as an out-of-the-box method for extracting a wide range of structural parameters with minimal user input.

Nina Weng, Paraskevas Pegios, Eike Petersen et al.

Shortcut learning is when a model -- e.g. a cardiac disease classifier -- exploits correlations between the target label and a spurious shortcut feature, e.g. a pacemaker, to predict the target label based on the shortcut rather than real discriminative features. This is common in medical imaging, where treatment and clinical annotations correlate with disease labels, making them easy shortcuts to predict disease. We propose a novel detection and quantification of the impact of potential shortcut features via a fast diffusion-based counterfactual image generation that can synthetically remove or add shortcuts. Via a novel inpainting-based modification we spatially limit the changes made with no extra inference step, encouraging the removal of spatially constrained shortcut features while ensuring that the shortcut-free counterfactuals preserve their remaining image features to a high degree. Using these, we assess how shortcut features influence model predictions. This is enabled by our second contribution: An efficient diffusion-based counterfactual explanation method with significant inference speed-up at comparable image quality as state-of-the-art. We confirm this on two large chest X-ray datasets, a skin lesion dataset, and CelebA. Our code is publicly available at fastdime.compute.dtu.dk.

Nina Weng, Aasa Feragen, Siavash Bigdeli

Diffusion-based generative models, such as Denoising Diffusion Probabilistic Models (DDPMs), have achieved remarkable success in image generation, but their step-by-step denoising process remains opaque, leaving critical aspects of the generation mechanism unexplained. To address this, we introduce \emph{Patronus}, an interpretable diffusion model inspired by ProtoPNet. Patronus integrates a prototypical network into DDPMs, enabling the extraction of prototypes and conditioning of the generation process on their prototype activation vector. This design enhances interpretability by showing the learned prototypes and how they influence the generation process. Additionally, the model supports downstream tasks like image manipulation, enabling more transparent and controlled modifications. Moreover, Patronus could reveal shortcut learning in the generation process by detecting unwanted correlations between learned prototypes. Notably, Patronus operates entirely without any annotations or text prompts. This work opens new avenues for understanding and controlling diffusion models through prototype-based interpretability. Our code is available at \href{https://github.com/nina-weng/patronus}{https://github.com/nina-weng/patronus}.

Lezhong Wang, Jeppe Revall Frisvad, Mark Bo Jensen et al.

The demand for stereo images increases as manufacturers launch more XR devices. To meet this demand, we introduce StereoDiffusion, a method that, unlike traditional inpainting pipelines, is trainning free, remarkably straightforward to use, and it seamlessly integrates into the original Stable Diffusion model. Our method modifies the latent variable to provide an end-to-end, lightweight capability for fast generation of stereo image pairs, without the need for fine-tuning model weights or any post-processing of images. Using the original input to generate a left image and estimate a disparity map for it, we generate the latent vector for the right image through Stereo Pixel Shift operations, complemented by Symmetric Pixel Shift Masking Denoise and Self-Attention Layers Modification methods to align the right-side image with the left-side image. Moreover, our proposed method maintains a high standard of image quality throughout the stereo generation process, achieving state-of-the-art scores in various quantitative evaluations.

Paraskevas Pegios, Manxi Lin, Nina Weng et al.

Obstetric ultrasound image quality is crucial for accurate diagnosis and monitoring of fetal health. However, acquiring high-quality standard planes is difficult, influenced by the sonographer's expertise and factors like the maternal BMI or fetus dynamics. In this work, we explore diffusion-based counterfactual explainable AI to generate realistic, high-quality standard planes from low-quality non-standard ones. Through quantitative and qualitative evaluation, we demonstrate the effectiveness of our approach in generating plausible counterfactuals of increased quality. This shows future promise for enhancing training of clinicians by providing visual feedback and potentially improving standard plane quality and acquisition for downstream diagnosis and monitoring.

Martina Mattioli, Eike Petersen, Aasa Feragen et al.

Artificial intelligence (AI) is being applied in almost every field. At the same time, the currently dominant deep learning methods are fundamentally black-box systems that lack explanations for their inferences, significantly limiting their trustworthiness and adoption. Explainable AI (XAI) methods aim to overcome this challenge by providing explanations of the models' decision process. Such methods are often proposed and developed by engineers and scientists with a predominantly technical background and incorporate their assumptions about the existence, validity, and explanatory utility of different conceivable explanatory mechanisms. However, the basic concept of an explanation -- what it is, whether we can know it, whether it is absolute or relative -- is far from trivial and has been the subject of deep philosophical debate for millennia. As we point out here, the assumptions incorporated into different XAI methods are not harmless and have important consequences for the validity and interpretation of AI explanations in different domains. We investigate ontological and epistemological assumptions in explainability methods when they are applied to AI systems, meaning the assumptions we make about the existence of explanations and our ability to gain knowledge about those explanations. Our analysis shows how seemingly small technical changes to an XAI method may correspond to important differences in the underlying assumptions about explanations. We furthermore highlight the risks of ignoring the underlying onto-epistemological paradigm when choosing an XAI method for a given application, and we discuss how to select and adapt appropriate XAI methods for different domains of application.

Lezhong Wang, Shutong Jin, Ruiqi Cui et al.

Controlling illumination during video post-production is a crucial yet elusive goal in computational photography. Existing methods often lack flexibility, restricting users to certain relighting models. This paper introduces ReLumix, a novel framework that decouples the relighting algorithm from temporal synthesis, thereby enabling any image relighting technique to be seamlessly applied to video. Our approach reformulates video relighting into a simple yet effective two-stage process: (1) an artist relights a single reference frame using any preferred image-based technique (e.g., Diffusion Models, physics-based renderers); and (2) a fine-tuned stable video diffusion (SVD) model seamlessly propagates this target illumination throughout the sequence. To ensure temporal coherence and prevent artifacts, we introduce a gated cross-attention mechanism for smooth feature blending and a temporal bootstrapping strategy that harnesses SVD's powerful motion priors. Although trained on synthetic data, ReLumix shows competitive generalization to real-world videos. The method demonstrates significant improvements in visual fidelity, offering a scalable and versatile solution for dynamic lighting control.

Lauritz Christian Holme, Anton Mosquera Storgaard, Siavash Arjomand Bigdeli

The rise of generative image models leads to privacy concerns when it comes to the huge datasets used to train such models. This paper investigates the possibility of inferring if a set of face images was used for fine-tuning a Latent Diffusion Model (LDM). A Membership Inference Attack (MIA) method is presented for this task. Using generated auxiliary data for the training of the attack model leads to significantly better performance, and so does the use of watermarks. The guidance scale used for inference was found to have a significant influence. If a LDM is fine-tuned for long enough, the text prompt used for inference has no significant influence. The proposed MIA is found to be viable in a realistic black-box setup against LDMs fine-tuned on face-images.

Pawel Tomasz Pieta, Peter Winkel Rasmussen, Anders Bjorholm Dahl et al.

Influenced by the complexity of volumetric imaging, there is a shortage of established datasets useful for benchmarking volumetric deep-learning models. As a consequence, new and existing models are not easily comparable, limiting the development of architectures optimized specifically for volumetric data. To counteract this trend, we introduce MozzaVID - a large, clean, and versatile volumetric classification dataset. Our dataset contains X-ray computed tomography (CT) images of mozzarella microstructure and enables the classification of 25 cheese types and 149 cheese samples. We provide data in three different resolutions, resulting in three dataset instances containing from 591 to 37,824 images. While being general-purpose, the dataset also facilitates investigating mozzarella structure properties. The structure of food directly affects its functional properties and thus its consumption experience. Understanding food structure helps tune the production and mimicking it enables sustainable alternatives to animal-derived food products. The complex and disordered nature of food structures brings a unique challenge, where a choice of appropriate imaging method, scale, and sample size is not trivial. With this dataset we aim to address these complexities, contributing to more robust structural analysis models. The dataset can be downloaded from: https://archive.compute.dtu.dk/files/public/projects/MozzaVID/.

Evangelos Ntavelis, Andrés Romero, Siavash Bigdeli et al.

This paper reviews the AIM 2020 challenge on extreme image inpainting. This report focuses on proposed solutions and results for two different tracks on extreme image inpainting: classical image inpainting and semantically guided image inpainting. The goal of track 1 is to inpaint considerably large part of the image using no supervision but the context. Similarly, the goal of track 2 is to inpaint the image by having access to the entire semantic segmentation map of the image to inpaint. The challenge had 88 and 74 participants, respectively. 11 and 6 teams competed in the final phase of the challenge, respectively. This report gauges current solutions and set a benchmark for future extreme image inpainting methods.

David Honzátko, Siavash A. Bigdeli, Engin Türetken et al.

Image denoising is an essential tool in computational photography. Standard denoising techniques, which use deep neural networks at their core, require pairs of clean and noisy images for its training. If we do not possess the clean samples, we can use blind-spot neural network architectures, which estimate the pixel value based on the neighbouring pixels only. These networks thus allow training on noisy images directly, as they by-design avoid trivial solutions. Nowadays, the blind-spot is mostly achieved using shifted convolutions or serialization. We propose a novel fully convolutional network architecture that uses dilations to achieve the blind-spot property. Our network improves the performance over the prior work and achieves state-of-the-art results on established datasets.

Tiziano Portenier, Siavash Bigdeli, Orcun Goksel

We present a novel texture synthesis framework, enabling the generation of infinite, high-quality 3D textures given a 2D exemplar image. Inspired by recent advances in natural texture synthesis, we train deep neural models to generate textures by non-linearly combining learned noise frequencies. To achieve a highly realistic output conditioned on an exemplar patch, we propose a novel loss function that combines ideas from both style transfer and generative adversarial networks. In particular, we train the synthesis network to match the Gram matrices of deep features from a discriminator network. In addition, we propose two architectural concepts and an extrapolation strategy that significantly improve generalization performance. In particular, we inject both model input and condition into hidden network layers by learning to scale and bias hidden activations. Quantitative and qualitative evaluations on a diverse set of exemplars motivate our design decisions and show that our system performs superior to previous state of the art. Finally, we conduct a user study that confirms the benefits of our framework.

Siavash A. Bigdeli, Geng Lin, Tiziano Portenier et al.

Learning probabilistic models that can estimate the density of a given set of samples, and generate samples from that density, is one of the fundamental challenges in unsupervised machine learning. We introduce a new generative model based on denoising density estimators (DDEs), which are scalar functions parameterized by neural networks, that are efficiently trained to represent kernel density estimators of the data. Leveraging DDEs, our main contribution is a novel technique to obtain generative models by minimizing the KL-divergence directly. We prove that our algorithm for obtaining generative models is guaranteed to converge to the correct solution. Our approach does not require specific network architecture as in normalizing flows, nor use ordinary differential equation solvers as in continuous normalizing flows. Experimental results demonstrate substantial improvement in density estimation and competitive performance in generative model training.

Siavash Bigdeli, David Honzátko, Sabine Süsstrunk et al.

Plug-and-play denoisers can be used to perform generic image restoration tasks independent of the degradation type. These methods build on the fact that the Maximum a Posteriori (MAP) optimization can be solved using smaller sub-problems, including a MAP denoising optimization. We present the first end-to-end approach to MAP estimation for image denoising using deep neural networks. We show that our method is guaranteed to minimize the MAP denoising objective, which is then used in an optimization algorithm for generic image restoration. We provide theoretical analysis of our approach and show the quantitative performance of our method in several experiments. Our experimental results show that the proposed method can achieve 70x faster performance compared to the state-of-the-art, while maintaining the theoretical perspective of MAP.

Edo Collins, Siavash Arjomand Bigdeli, Sabine Süsstrunk

We propose a new notion of `non-linearity' of a network layer with respect to an input batch that is based on its proximity to a linear system, which is reflected in the non-negative rank of the activation matrix. We measure this non-linearity by applying non-negative factorization to the activation matrix. Considering batches of similar samples, we find that high non-linearity in deep layers is indicative of memorization. Furthermore, by applying our approach layer-by-layer, we find that the mechanism for memorization consists of distinct phases. We perform experiments on fully-connected and convolutional neural networks trained on several image and audio datasets. Our results demonstrate that as an indicator for memorization, our technique can be used to perform early stopping.

Tiziano Portenier, Qiyang Hu, Attila Szabó et al.

We present a novel system for sketch-based face image editing, enabling users to edit images intuitively by sketching a few strokes on a region of interest. Our interface features tools to express a desired image manipulation by providing both geometry and color constraints as user-drawn strokes. As an alternative to the direct user input, our proposed system naturally supports a copy-paste mode, which allows users to edit a given image region by using parts of another exemplar image without the need of hand-drawn sketching at all. The proposed interface runs in real-time and facilitates an interactive and iterative workflow to quickly express the intended edits. Our system is based on a novel sketch domain and a convolutional neural network trained end-to-end to automatically learn to render image regions corresponding to the input strokes. To achieve high quality and semantically consistent results we train our neural network on two simultaneous tasks, namely image completion and image translation. To the best of our knowledge, we are the first to combine these two tasks in a unified framework for interactive image editing. Our results show that the proposed sketch domain, network architecture, and training procedure generalize well to real user input and enable high quality synthesis results without additional post-processing.

Siavash Arjomand Bigdeli, Meiguang Jin, Paolo Favaro et al.

In this paper we introduce a natural image prior that directly represents a Gaussian-smoothed version of the natural image distribution. We include our prior in a formulation of image restoration as a Bayes estimator that also allows us to solve noise-blind image restoration problems. We show that the gradient of our prior corresponds to the mean-shift vector on the natural image distribution. In addition, we learn the mean-shift vector field using denoising autoencoders, and use it in a gradient descent approach to perform Bayes risk minimization. We demonstrate competitive results for noise-blind deblurring, super-resolution, and demosaicing.

Siavash Arjomand Bigdeli, Matthias Zwicker

We propose to leverage denoising autoencoder networks as priors to address image restoration problems. We build on the key observation that the output of an optimal denoising autoencoder is a local mean of the true data density, and the autoencoder error (the difference between the output and input of the trained autoencoder) is a mean shift vector. We use the magnitude of this mean shift vector, that is, the distance to the local mean, as the negative log likelihood of our natural image prior. For image restoration, we maximize the likelihood using gradient descent by backpropagating the autoencoder error. A key advantage of our approach is that we do not need to train separate networks for different image restoration tasks, such as non-blind deconvolution with different kernels, or super-resolution at different magnification factors. We demonstrate state of the art results for non-blind deconvolution and super-resolution using the same autoencoding prior.