Image registration is a fundamental medical image analysis task, and a wide variety of approaches have been proposed. However, only a few studies have comprehensively compared medical image registration approaches on a wide range of clinically relevant tasks. This limits the development of registration methods, the adoption of research advances into practice, and a fair benchmark across competing approaches. The Learn2Reg challenge addresses these limitations by providing a multi-task medical image registration data set for comprehensive characterisation of deformable registration algorithms. A continuous evaluation will be possible at https://learn2reg.grand-challenge.org. Learn2Reg covers a wide range of anatomies (brain, abdomen, and thorax), modalities (ultrasound, CT, MR), availability of annotations, as well as intra- and inter-patient registration evaluation. We established an easily accessible framework for training and validation of 3D registration methods, which enabled the compilation of results of over 65 individual method submissions from more than 20 unique teams. We used a complementary set of metrics, including robustness, accuracy, plausibility, and runtime, enabling unique insight into the current state-of-the-art of medical image registration. This paper describes datasets, tasks, evaluation methods and results of the challenge, as well as results of further analysis of transferability to new datasets, the importance of label supervision, and resulting bias. While no single approach worked best across all tasks, many methodological aspects could be identified that push the performance of medical image registration to new state-of-the-art performance. Furthermore, we demystified the common belief that conventional registration methods have to be much slower than deep-learning-based methods.

Vision Transformers (ViT)-based models are witnessing an exponential growth in the medical imaging community. Among desirable properties, ViTs provide a powerful modeling of long-range pixel relationships, contrary to inherently local convolutional neural networks (CNN). These emerging models can be categorized either as hybrid-based when used in conjunction with CNN layers (CNN-ViT) or purely Transformers-based. In this work, we conduct a comparative quantitative analysis to study the differences between a range of available Transformers-based models using controlled brain tumor segmentation experiments. We also investigate to what extent such models could benefit from modality interaction schemes in a multi-modal setting. Results on the publicly-available BraTS2021 dataset show that hybrid-based pipelines generally tend to outperform simple Transformers-based models. In these experiments, no particular improvement using multi-modal interaction schemes was observed.

We propose a new method for the progressive enhancement of images degraded by noise and blur using shock filters. The originality of our approach lies in the iterative exploitation of nonconservative edge-based force fields normally designed for segmentation to determine both locations and amplitudes of the sharpening action. The proposed nonconservative shock filter (NCSF) produces images with strong discontinuities that can be used as an additional pre-processing step to facilitate higher level tasks such as edge detection or segmentation. NCSF is especially useful when coupled with an anisotropic diffusion equation in noisy configurations. Using the proposed method, edges that are not well defined prior to filtering can be iteratively recovered. We show processing results on various types of images, focusing on medical transrectal ultrasound and brain PET imaging.

We propose to consider linear and nonlinear image diffusion processes as the result of the transport of image values in external advection fields. Linear diffusion is seen as the continuous transport of particles in fields of random orientations, while non-linear behavior is obtained by modulating the magnitude of these fields according to local edge strength. Edge sharpening can be achieved by biasing fields towards image edges using e.g. Gradient Vector Flow. This point of view, closely related to discrete random walks and other stochastic interpretations of diffusion processes, allows for simple, stable and surprisingly efficient filters that can be implemented using a few lines of code only.

In this paper, we extend the gradient vector flow field to the vector-valued case for robust variational segmentation of 4D images with active surfaces. Instead of only exploiting scalar edge strength in order to identify vector edges, we propagate both directions and amplitudes of vector gradients computed from the analysis of a structure tensor of the vector-valued image. To reduce contributions from noise in the calculation of the structure tensor, image channels are weighted according to a blind estimator of contrast that take profit of the deformable models framework. The proposed 4DGVF vector field is validated on synthetic image datasets and applied to biological volume delineation in dynamic PET imaging.

In this paper, we propose a new method for vector-valued image restoration in order to reduce noise while simultaneously sharpening vector edges. Our approach is a coupled anisotropic diffusion and shock filtering scheme that exploits a new robust 4DGVF vector field tailored for vector-valued images. The proposed scheme sharpens edges in directions diffused from the entire spatio-spectral information available with a more precise and a more stable sharpening effect along the iterative processing. We validate our method on color images as well as on realistic simulations of dynamic PET images.

In this paper, we extend the gradient vector flow field for robust variational segmentation of vector-valued images. Rather than using scalar edge information, we define a vectorial edge map derived from a weighted local structure tensor of the image that enables the diffusion of the gradient vectors in accurate directions through the 4D gradient vector flow equation. To reduce the contribution of noise in the structure tensor, image channels are weighted according to a blind estimator of contrast. The method is applied to biological volume delineation in dynamic PET imaging, and validated on realistic Monte Carlo simulations of numerical phantoms as well as on real images.