UNet 3+: A Full-Scale Connected UNet for Medical Image Segmentation

TL;DR

UNet 3+ introduces full-scale skip connections and deep supervision to enhance medical image segmentation accuracy and efficiency, especially for organs at varying scales, outperforming previous models.

Contribution

It proposes a novel UNet 3+ architecture with full-scale skip connections and deep supervision, improving segmentation accuracy and reducing parameters compared to prior models.

Findings

Improved segmentation accuracy on two datasets.

Reduced network parameters for better efficiency.

Enhanced boundary detection and reduced over-segmentation.

Abstract

Recently, a growing interest has been seen in deep learning-based semantic segmentation. UNet, which is one of deep learning networks with an encoder-decoder architecture, is widely used in medical image segmentation. Combining multi-scale features is one of important factors for accurate segmentation. UNet++ was developed as a modified Unet by designing an architecture with nested and dense skip connections. However, it does not explore sufficient information from full scales and there is still a large room for improvement. In this paper, we propose a novel UNet 3+, which takes advantage of full-scale skip connections and deep supervisions. The full-scale skip connections incorporate low-level details with high-level semantics from feature maps in different scales; while the deep supervision learns hierarchical representations from the full-scale aggregated feature maps. The proposed method is especially benefiting for organs that appear at varying scales. In addition to accuracy improvements, the proposed UNet 3+ can reduce the network parameters to improve the computation efficiency. We further propose a hybrid loss function and devise a classification-guided module to enhance the organ boundary and reduce the over-segmentation in a non-organ image, yielding more accurate segmentation results. The effectiveness of the proposed method is demonstrated on two datasets. The code is available at: github.com/ZJUGiveLab/UNet-Version