Semi-overcomplete convolutional auto-encoder embedding as shape priors for deep vessel segmentation
This addresses the challenge of delineating vascular systems in medical images for clinicians in diagnosis and planning, with incremental improvements over existing methods.
The paper tackles the problem of automatic vessel segmentation in medical images by integrating shape priors from a Semi-Overcomplete Convolutional Auto-Encoder (S-OCAE) embedding into deep segmentation models, showing effectiveness on retinal and liver vessel datasets compared to U-Net without priors and with traditional CAE priors.
The extraction of blood vessels has recently experienced a widespread interest in medical image analysis. Automatic vessel segmentation is highly desirable to guide clinicians in computer-assisted diagnosis, therapy or surgical planning. Despite a good ability to extract large anatomical structures, the capacity of U-Net inspired architectures to automatically delineate vascular systems remains a major issue, especially given the scarcity of existing datasets. In this paper, we present a novel approach that integrates into deep segmentation shape priors from a Semi-Overcomplete Convolutional Auto-Encoder (S-OCAE) embedding. Compared to standard Convolutional Auto-Encoders (CAE), it exploits an over-complete branch that projects data onto higher dimensions to better characterize tiny structures. Experiments on retinal and liver vessel extraction, respectively performed on publicly-available DRIVE and 3D-IRCADb datasets, highlight the effectiveness of our method compared to U-Net trained without and with shape priors from a traditional CAE.