DALight-3D: A Lightweight 3D U-Net for Brain Tumor Segmentation from Multi-Modal MRI
Nand Kumar Mishra, Dhruv Mishra, Dr Manu Pratap Singh
TLDR
DALight-3D is a lightweight 3D U-Net for brain tumor segmentation, achieving better accuracy-efficiency than baselines.
Key contributions
- Introduces DALight-3D, a compact 3D U-Net for efficient brain tumor segmentation.
- Integrates depthwise separable 3D convolutions, identifier-conditioned normalization, and cross-slice attention.
- Achieves a Dice score of 0.727 with 2.22M parameters, outperforming Residual 3D U-Net.
Why it matters
Volumetric brain tumor segmentation models are often computationally expensive. This paper addresses this by proposing DALight-3D, a lightweight 3D U-Net that significantly reduces parameters while maintaining or improving accuracy. Its efficient design makes advanced medical image analysis more accessible for clinical applications.
Original 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.
📬 Weekly AI Paper Digest
Get the top 10 AI/ML arXiv papers from the week — summarized, scored, and delivered to your inbox every Monday.