Directional Connectivity-based Segmentation of Medical Images
This work addresses the need for more accurate and consistent segmentation in medical image analysis, which is crucial for tasks like biomarker identification, but it appears incremental as it builds on existing connectivity-based approaches by adding directional information.
The paper tackles the problem of achieving anatomically consistent segmentation in medical images by incorporating directional connectivity modeling, resulting in enhanced feature representation and improved performance on public benchmarks compared to state-of-the-art methods.
Anatomical consistency in biomarker segmentation is crucial for many medical image analysis tasks. A promising paradigm for achieving anatomically consistent segmentation via deep networks is incorporating pixel connectivity, a basic concept in digital topology, to model inter-pixel relationships. However, previous works on connectivity modeling have ignored the rich channel-wise directional information in the latent space. In this work, we demonstrate that effective disentanglement of directional sub-space from the shared latent space can significantly enhance the feature representation in the connectivity-based network. To this end, we propose a directional connectivity modeling scheme for segmentation that decouples, tracks, and utilizes the directional information across the network. Experiments on various public medical image segmentation benchmarks show the effectiveness of our model as compared to the state-of-the-art methods. Code is available at https://github.com/Zyun-Y/DconnNet.