Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Semi-overcomplete convolutional auto-encoder embedding as shape priors for deep vessel segmentation

Amine Sadikine, Bogdan Badic, Jean-Pierre Tasu, Vincent Noblet, Dimitris Visvikis, Pierre-Henri Conze

TL;DR

A novel approach that integrates into deep segmentation shape priors from a Semi-Overcomplete Convolutional Auto-Encoder (S-OCAE) embedding, and exploits an over-complete branch that projects data onto higher dimensions to better characterize tiny structures.

Abstract

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.

Semi-overcomplete convolutional auto-encoder embedding as shape priors for deep vessel segmentation

TL;DR

A novel approach that integrates into deep segmentation shape priors from a Semi-Overcomplete Convolutional Auto-Encoder (S-OCAE) embedding, and exploits an over-complete branch that projects data onto higher dimensions to better characterize tiny structures.

Abstract

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.
Paper Structure (10 sections, 5 equations, 4 figures, 1 table)

This paper contains 10 sections, 5 equations, 4 figures, 1 table.

Table of Contents

  1. Introduction
  2. Methods
  3. Standard undercomplete shape priors
  4. Extension to semi-overcomplete shape priors
  5. Experiments
  6. Imaging datasets
  7. Implementation details
  8. Evaluation of predicted segmentation
  9. Results and discussion
  10. Conclusion

Figures (4)

  • Figure 1: Proposed pipeline. U-Net parameters are estimated by penalizing a segmentation loss $\ell_{\phi}$ as well as a regularization term $\ell_{s}$ dealing with the similarity between projections of prediction and ground truth in a learned S-OCAE latent space.
  • Figure 2: Proposed S-OCAE network whose multi-path encoder is made of undercomplete and overcomplete branches and includes communication (CB), fusion (FB) and residual blocks.
  • Figure 3: Liver vessel segmentation results. Ground truth and predicted contours are respectively in green and blue.
  • Figure 4: Retinal vessel segmentation results. True positives, false positives and false negatives are in white, red and yellow.