DAUNet: A Lightweight UNet Variant with Deformable Convolutions and Parameter-Free Attention for Medical Image Segmentation

TL;DR

DAUNet is a lightweight, efficient medical image segmentation model that combines deformable convolutions and parameter-free attention to improve accuracy and robustness without increasing complexity.

Contribution

It introduces a novel lightweight UNet variant integrating deformable V2 convolutions and SimAM attention, enhancing spatial adaptability and feature fusion in medical imaging.

Findings

Outperforms state-of-the-art models in Dice score, HD95, and ASD.

Maintains high parameter efficiency and robustness in low-contrast regions.

Effective in real-time, resource-constrained clinical environments.

Abstract

Medical image segmentation plays a pivotal role in automated diagnostic and treatment planning systems. In this work, we present DAUNet, a novel lightweight UNet variant that integrates Deformable V2 Convolutions and Parameter-Free Attention (SimAM) to improve spatial adaptability and context-aware feature fusion without increasing model complexity. DAUNet's bottleneck employs dynamic deformable kernels to handle geometric variations, while the decoder and skip pathways are enhanced using SimAM attention modules for saliency-aware refinement. Extensive evaluations on two challenging datasets, FH-PS-AoP (fetal head and pubic symphysis ultrasound) and FUMPE (CT-based pulmonary embolism detection), demonstrate that DAUNet outperforms state-of-the-art models in Dice score, HD95, and ASD, while maintaining superior parameter efficiency. Ablation studies highlight the individual contributions of deformable convolutions and SimAM attention. DAUNet's robustness to missing context and low-contrast regions establishes its suitability for deployment in real-time and resource-constrained clinical environments.