Deep learning models often need sufficient supervision (i.e., labelled data) in order to be trained effectively. By contrast, humans can swiftly learn to identify important anatomy in medical images like MRI and CT scans, with minimal guidance. This recognition capability easily generalises to new images from different medical facilities and to new tasks in different settings. This rapid and generalisable learning ability is largely due to the compositional structure of image patterns in the human brain, which are not well represented in current medical models. In this paper, we study the utilisation of compositionality in learning more interpretable and generalisable representations for medical image segmentation. Overall, we propose that the underlying generative factors that are used to generate the medical images satisfy compositional equivariance property, where each factor is compositional (e.g., corresponds to human anatomy) and also equivariant to the task. Hence, a good representation that approximates well the ground truth factor has to be compositionally equivariant. By modelling the compositional representations with learnable von-Mises-Fisher (vMF) kernels, we explore how different design and learning biases can be used to enforce the representations to be more compositionally equivariant under un-, weakly-, and semi-supervised settings. Extensive results show that our methods achieve the best performance over several strong baselines on the task of semi-supervised domain-generalised medical image segmentation. Code will be made publicly available upon acceptance at https://github.com/vios-s.

Traditional marker-based motion capture requires excessive and specialized equipment, hindering accessibility and wider adoption. In this work, we demonstrate such a system but rely on a very sparse set of low-cost consumer-grade sensors. Our system exploits a data-driven backend to infer the captured subject's joint positions from noisy marker estimates in real-time. In addition to reduced costs and portability, its inherent denoising nature allows for quicker captures by alleviating the need for precise marker placement and post-processing, making it suitable for interactive virtual reality applications.

Disentangled representation learning has been proposed as an approach to learning general representations even in the absence of, or with limited, supervision. A good general representation can be fine-tuned for new target tasks using modest amounts of data, or used directly in unseen domains achieving remarkable performance in the corresponding task. This alleviation of the data and annotation requirements offers tantalising prospects for applications in computer vision and healthcare. In this tutorial paper, we motivate the need for disentangled representations, revisit key concepts, and describe practical building blocks and criteria for learning such representations. We survey applications in medical imaging emphasising choices made in exemplar key works, and then discuss links to computer vision applications. We conclude by presenting limitations, challenges, and opportunities.

Learning to understand and infer object functionalities is an important step towards robust visual intelligence. Significant research efforts have recently focused on segmenting the object parts that enable specific types of human-object interaction, the so-called "object affordances". However, most works treat it as a static semantic segmentation problem, focusing solely on object appearance and relying on strong supervision and object detection. In this paper, we propose a novel approach that exploits the spatio-temporal nature of human-object interaction for affordance segmentation. In particular, we design an autoencoder that is trained using ground-truth labels of only the last frame of the sequence, and is able to infer pixel-wise affordance labels in both videos and static images. Our model surpasses the need for object labels and bounding boxes by using a soft-attention mechanism that enables the implicit localization of the interaction hotspot. For evaluation purposes, we introduce the SOR3D-AFF corpus, which consists of human-object interaction sequences and supports 9 types of affordances in terms of pixel-wise annotation, covering typical manipulations of tool-like objects. We show that our model achieves competitive results compared to strongly supervised methods on SOR3D-AFF, while being able to predict affordances for similar unseen objects in two affordance image-only datasets.

With the advent of consumer grade depth sensors, low-cost volumetric capture systems are easier to deploy. Their wider adoption though depends on their usability and by extension on the practicality of spatially aligning multiple sensors. Most existing alignment approaches employ visual patterns, e.g. checkerboards, or markers and require high user involvement and technical knowledge. More user-friendly and easier-to-use approaches rely on markerless methods that exploit geometric patterns of a physical structure. However, current SoA approaches are bounded by restrictions in the placement and the number of sensors. In this work, we improve markerless data-driven correspondence estimation to achieve more robust and flexible multi-sensor spatial alignment. In particular, we incorporate geometric constraints in an end-to-end manner into a typical segmentation based model and bridge the intermediate dense classification task with the targeted pose estimation one. This is accomplished by a soft, differentiable procrustes analysis that regularizes the segmentation and achieves higher extrinsic calibration performance in expanded sensor placement configurations, while being unrestricted by the number of sensors of the volumetric capture system. Our model is experimentally shown to achieve similar results with marker-based methods and outperform the mark-erless ones, while also being robust to the pose variations of the calibration structure. Code and pretrained models are available at https://vcl3d.github.io/StructureNet/.

Depth perception is considered an invaluable source of information for various vision tasks. However, depth maps acquired using consumer-level sensors still suffer from non-negligible noise. This fact has recently motivated researchers to exploit traditional filters, as well as the deep learning paradigm, in order to suppress the aforementioned non-uniform noise, while preserving geometric details. Despite the effort, deep depth denoising is still an open challenge mainly due to the lack of clean data that could be used as ground truth. In this paper, we propose a fully convolutional deep autoencoder that learns to denoise depth maps, surpassing the lack of ground truth data. Specifically, the proposed autoencoder exploits multiple views of the same scene from different points of view in order to learn to suppress noise in a self-supervised end-to-end manner using depth and color information during training, yet only depth during inference. To enforce self-supervision, we leverage a differentiable rendering technique to exploit photometric supervision, which is further regularized using geometric and surface priors. As the proposed approach relies on raw data acquisition, a large RGB-D corpus is collected using Intel RealSense sensors. Complementary to a quantitative evaluation, we demonstrate the effectiveness of the proposed self-supervised denoising approach on established 3D reconstruction applications. Code is avalable at https://github.com/VCL3D/DeepDepthDenoising.

Sensorimotor learning, namely the process of understanding the physical world by combining visual and motor information, has been recently investigated, achieving promising results for the task of 2D/3D object recognition. Following the recent trend in computer vision, powerful deep neural networks (NNs) have been used to model the “sensory” and “motor” information, namely the object appearance and affordance. However, the existing implementations cannot efficiently address the spatio-temporal nature of the human-object interaction. Inspired by recent work on attention-based learning, this paper introduces an attention-enhanced NN-based model that learns to selectively focus on parts of the physical interaction where the object appearance is corrupted by occlusions and deformations. The model's attention mechanism relies on the confidence of classifying an object based solely on its appearance. Three metrics are used to measure the latter, namely the prediction entropy, the average N-best likelihood difference, and the N-best likelihood dispersion. Evaluation of the attention-enhanced model on the SOR3D dataset reports 33% and 26% relative improvement over the appearance-only and the spatio-temporal fusion baseline models, respectively.

It is well-established by cognitive neuroscience that human perception of objects constitutes a complex process, where object appearance information is combined with evidence about the so-called object affordances, namely the types of actions that humans typically perform when interacting with them. This fact has recently motivated the sensorimotor approach to the challenging task of automatic object recognition, where both information sources are fused to improve robustness. In this work, the aforementioned paradigm is adopted, surpassing current limitations of sensorimotor object recognition research. Specifically, the deep learning paradigm is introduced to the problem for the first time, developing a number of novel neuro-biologically and neuro-physiologically inspired architectures that utilize state-of-the-art neural networks for fusing the available information sources in multiple ways. The proposed methods are evaluated using a large RGB-D corpus, which is specifically collected for the task of sensorimotor object recognition and is made publicly available. Experimental results demonstrate the utility of affordance information to object recognition, achieving an up to 29% relative error reduction by its inclusion.

Motivated by increasing popularity of depth visual sensors, such as the Kinect device, we investigate the utility of depth information in audio-visual speech activity detection. A two-subject scenario is assumed, allowing to also consider speech overlap. Two sensory setups are employed, where depth video captures either a frontal or profile view of the subjects, and is subsequently combined with the corresponding planar video and audio streams. Further, multi-view fusion is regarded, using audio and planar video from a sensor at the complementary view setup. Support vector machines provide temporal speech activity classification for each visually detected subject, fusing the available modality streams. Classification results are further combined to yield speaker diarization. Experiments are reported on a suitable audio-visual corpus recorded by two Kinects. Results demonstrate the benefits of depth information, particularly in the frontal depth view setup, reducing speech activity detection and speaker diarization errors over systems that ignore it.