LightMedSeg: Lightweight 3D Medical Image Segmentation with Learned Spatial Anchors
This work provides a deployable and data-efficient solution for clinicians and researchers needing to perform 3D medical image segmentation in resource-constrained environments.
This paper addresses the challenge of accurate and efficient 3D medical image segmentation under tight computational constraints. The proposed LightMedSeg model, with only 0.48M parameters and 14.64 GFLOPs, achieves segmentation accuracy comparable to much larger transformer-based models, differing by only a few Dice points.
Accurate and efficient 3D medical image segmentation is essential for clinical AI, where models must remain reliable under stringent memory, latency, and data availability constraints. Transformer-based methods achieve strong accuracy but suffer from excessive parameters, high FLOPs, and limited generalization. We propose LightMedSeg, a modular UNet-style segmentation architecture that integrates anatomical priors with adaptive context modeling. Anchor-conditioned FiLM modulation enables anatomy-aware feature calibration, while a local structural prior module and texture-aware routing dynamically allocate representational capacity to boundary-rich regions. Computational redundancy is minimized through ghost and depthwise convolutions, and multi-scale features are adaptively fused via a learned skip router with anchor-relative spatial position bias. Despite requiring only 0.48M parameters and 14.64~GFLOPs, LightMedSeg achieves segmentation accuracy within a few Dice points of heavy transformer baselines. Therefore, LightMedSeg is a deployable and data-efficient solution for 3D medical image segmentation. Code will be released publicly upon acceptance.