The Importance of Skip Connections in Biomedical Image Segmentation

TL;DR

This paper investigates the impact of both long and short skip connections in deep Fully Convolutional Networks for biomedical image segmentation, demonstrating that their combination improves gradient flow and segmentation performance.

Contribution

It introduces the use of short skip connections in very deep FCNs for biomedical segmentation, enhancing gradient flow and achieving near-state-of-the-art results.

Findings

Short skip connections improve gradient flow in deep FCNs.

Deep FCNs with combined skip connections outperform standard models.

Near state-of-the-art segmentation results on EM dataset without post-processing.

Abstract

In this paper, we study the influence of both long and short skip connections on Fully Convolutional Networks (FCN) for biomedical image segmentation. In standard FCNs, only long skip connections are used to skip features from the contracting path to the expanding path in order to recover spatial information lost during downsampling. We extend FCNs by adding short skip connections, that are similar to the ones introduced in residual networks, in order to build very deep FCNs (of hundreds of layers). A review of the gradient flow confirms that for a very deep FCN it is beneficial to have both long and short skip connections. Finally, we show that a very deep FCN can achieve near-to-state-of-the-art results on the EM dataset without any further post-processing.