SIB-UNet: A dual encoder medical image segmentation model with selective fusion and information bottleneck fusion

Published: 01 Jan 2024, Last Modified: 11 Apr 2025Expert Syst. Appl. 2024EveryoneRevisionsBibTeXCC BY-SA 4.0
Abstract: Highlights•Dual encoder structure, including selective fusion module, can dynamically fuse global and local features.•The skip connection structure based on information bottleneck theory can narrow the semantic gap.•Obtain better segmentation results with less resource consumption.
Loading