DALight-3D: A Lightweight 3D U-Net for Brain Tumor Segmentation from Multi-Modal MRI

TL;DR

DALight-3D introduces a lightweight 3D U-Net variant optimized for brain tumor segmentation from multi-modal MRI, achieving high accuracy with fewer parameters and lower computational cost.

Contribution

The paper presents DALight-3D, a novel compact 3D U-Net architecture combining advanced convolutional and normalization techniques for efficient brain tumor segmentation.

Findings

DALight-3D achieves a mean Dice of 0.727 with 2.22M parameters.

It outperforms standard 3D U-Net variants in accuracy and efficiency.

Component ablations confirm the importance of each proposed module.

Abstract

Automatic brain tumor segmentation from multi-modal MRI remains challenging because volumetric models often incur substantial computational cost. This paper presents DALight-3D, a compact 3D U-Net variant that combines depthwise separable 3D convolutions, identifier-conditioned normalization, cross-slice attention, and adaptive skip fusion. The method is evaluated on the Medical Segmentation Decathlon Task01 BrainTumour benchmark under matched optimization settings against standard 3D U-Net, Attention U-Net, Residual 3D U-Net, and V-Net baselines. In the reported 50-epoch comparison, DALight-3D achieves a mean Dice of 0.727 with 2.22M parameters, compared with 0.710 Dice and 3.20M parameters for Residual 3D U-Net. Component-wise ablations show consistent performance degradation when SepConv, identifier-conditioned normalization, CSA, or SSFB is removed. These results indicate that DALight-3D offers a favorable accuracy-efficiency trade-off within the present benchmark setting.