State-of-the-Art Stroke Lesion Segmentation at 1/1000th of Parameters
This work addresses the problem of resource-intensive medical image analysis for clinicians and researchers in resource-limited environments, offering an incremental improvement in efficiency.
The paper tackled efficient whole-brain stroke lesion segmentation by revisiting MeshNet with a novel multi-scale dilation pattern and encoder-decoder structure, achieving superior or comparable DICE scores to state-of-the-art models at 1/1000th of the parameters on the ARC dataset.
Efficient and accurate whole-brain lesion segmentation remains a challenge in medical image analysis. In this work, we revisit MeshNet, a parameter-efficient segmentation model, and introduce a novel multi-scale dilation pattern with an encoder-decoder structure. This innovation enables capturing broad contextual information and fine-grained details without traditional downsampling, upsampling, or skip-connections. Unlike previous approaches processing subvolumes or slices, we operate directly on whole-brain $256^3$ MRI volumes. Evaluations on the Aphasia Recovery Cohort (ARC) dataset demonstrate that MeshNet achieves superior or comparable DICE scores to state-of-the-art architectures such as MedNeXt and U-MAMBA at 1/1000th of parameters. Our results validate MeshNet's strong balance of efficiency and performance, making it particularly suitable for resource-limited environments such as web-based applications and opening new possibilities for the widespread deployment of advanced medical image analysis tools.