Two Views Are Better than One: Monocular 3D Pose Estimation with Multiview Consistency

Deducing a 3D human pose from a single 2D image or 2D keypoints is inherently challenging, given the fundamental ambiguity wherein multiple 3D poses can correspond to the same 2D representation. The acquisition of 3D data, while invaluable for resolving pose ambiguity, is expensive and requires an intricate setup, often restricting its applicability to controlled lab environments. We improve performance of monocular human pose estimation models using multiview data for fine-tuning. We propose a novel loss function, multiview consistency, to enable adding additional training data with only 2D supervision. This loss enforces that the inferred 3D pose from one view aligns with the inferred 3D pose from another view under similarity transformations. Our consistency loss substantially improves performance for fine-tuning with no available 3D data. Our experiments demonstrate that two views offset by 90 degrees are enough to obtain good performance, with only marginal improvements by adding more views. Thus, we enable the acquisition of domain-specific data by capturing activities with off-the-shelf cameras, eliminating the need for elaborate calibration procedures. This research introduces new possibilities for domain adaptation in 3D pose estimation, providing a practical and cost-effective solution to customize models for specific applications. The used dataset, featuring additional views, will be made publicly available.

SportsPose -- A Dynamic 3D sports pose dataset

Accurate 3D human pose estimation is essential for sports analytics, coaching, and injury prevention. However, existing datasets for monocular pose estimation do not adequately capture the challenging and dynamic nature of sports movements. In response, we introduce SportsPose, a large-scale 3D human pose dataset consisting of highly dynamic sports movements. With more than 176,000 3D poses from 24 different subjects performing 5 different sports activities, SportsPose provides a diverse and comprehensive set of 3D poses that reflect the complex and dynamic nature of sports movements. Contrary to other markerless datasets we have quantitatively evaluated the precision of SportsPose by comparing our poses with a commercial marker-based system and achieve a mean error of 34.5 mm across all evaluation sequences. This is comparable to the error reported on the commonly used 3DPW dataset. We further introduce a new metric, local movement, which describes the movement of the wrist and ankle joints in relation to the body. With this, we show that SportsPose contains more movement than the Human3.6M and 3DPW datasets in these extremum joints, indicating that our movements are more dynamic. The dataset with accompanying code can be downloaded from our website. We hope that SportsPose will allow researchers and practitioners to develop and evaluate more effective models for the analysis of sports performance and injury prevention. With its realistic and diverse dataset, SportsPose provides a valuable resource for advancing the state-of-the-art in pose estimation in sports.

BugNIST -- A New Large Scale Volumetric 3D Image Dataset for Classification and Detection

Progress in 3D volumetric image analysis research is limited by the lack of datasets and most advances in analysis methods for volumetric images are based on medical data. However, medical data do not necessarily resemble the characteristics of other volumetric images such as micro-CT. To promote research in 3D volumetric image analysis beyond medical data, we have created the BugNIST dataset and made it freely available. BugNIST is an extensive dataset of micro-CT scans of 12 types of bugs, such as insects and larvae. BugNIST contains 9437 volumes where 9087 are of individual bugs and 350 are mixtures of bugs and other material. The goal of BugNIST is to benchmark classification and detection methods, and we have designed the detection challenge such that detection models are trained on scans of individual bugs and tested on bug mixtures. Models capable of solving this task will be independent of the context, i.e., the surrounding material. This is a great advantage if the context is unknown or changing, as is often the case in micro-CT. Our initial baseline analysis shows that current state-of-the-art deep learning methods classify individual bugs very well, but has great difficulty with the detection challenge. Hereby, BugNIST enables research in image analysis areas that until now have missed relevant data - both classification, detection, and hopefully more.

Sparse-View Spectral CT Reconstruction Using Deep Learning

Spectral CT is an emerging technology capable of providing high chemical specificity, which is crucial for many applications such as detecting threats in luggage. Such applications often require both fast and high-quality image reconstruction based on sparse-view (few) projections. The conventional FBP method is fast but it produces low-quality images dominated by noise and artifacts when few projections are available. Iterative methods with, e.g., TV regularizers can circumvent that but they are computationally expensive, with the computational load proportionally increasing with the number of spectral channels. Instead, we propose an approach for fast reconstruction of sparse-view spectral CT data using U-Net with multi-channel input and output. The network is trained to output high-quality images from input images reconstructed by FBP. The network is fast at run-time and because the internal convolutions are shared between the channels, the computation load increases only at the first and last layers, making it an efficient approach to process spectral data with a large number of channels. We validated our approach using real CT scans. The results show qualitatively and quantitatively that our approach is able to outperform the state-of-the-art iterative methods. Furthermore, the results indicate that the network is able to exploit the coupling of the channels to enhance the overall quality and robustness.

Multi-Spectral Imaging via Computed Tomography (MUSIC) - Comparing Unsupervised Spectral Segmentations for Material Differentiation

Multi-spectral computed tomography is an emerging technology for the non-destructive identification of object materials and the study of their physical properties. Applications of this technology can be found in various scientific and industrial contexts, such as luggage scanning at airports. Material distinction and its identification is challenging, even with spectral x-ray information, due to acquisition noise, tomographic reconstruction artefacts and scanning setup application constraints. We present MUSIC - and open access multi-spectral CT dataset in 2D and 3D - to promote further research in the area of material identification. We demonstrate the value of this dataset on the image analysis challenge of object segmentation purely based on the spectral response of its composing materials. In this context, we compare the segmentation accuracy of fast adaptive mean shift (FAMS) and unconstrained graph cuts on both datasets. We further discuss the impact of reconstruction artefacts and segmentation controls on the achievable results. Dataset, related software packages and further documentation are made available to the imaging community in an open-access manner to promote further data-driven research on the subject

Content-based Propagation of User Markings for Interactive Segmentation of Patterned Images

Efficient and easy segmentation of images and volumes is of great practical importance. Segmentation problems which motivate our approach originate from imaging commonly used in materials science and medicine. We formulate image segmentation as a probabilistic pixel classification problem, and we apply segmentation as a step towards characterising image content. Our method allows the user to define structures of interest by interactively marking a subset of pixels. Thanks to the real-time feedback, the user can place new markings strategically, depending on the current outcome. The final pixel classification may be obtained from a very modest user input. An important ingredient of our method is a graph that encodes image content. This graph is built in an unsupervised manner during initialisation, and is based on clustering of image features. Since we combine a limited amount of user-labelled data with the clustering information obtained from the unlabelled parts of the image, our method fits in the general framework of semi-supervised learning. We demonstrate how this can be a very efficient approach to segmentation through pixel classification.