GCtx-UNet: Efficient Network for Medical Image Segmentation

TL;DR

GCtx-UNet is a lightweight, efficient medical image segmentation network that combines global and local features using vision transformers, outperforming existing models in accuracy and computational efficiency across multiple datasets.

Contribution

The paper introduces GCtx-UNet, a novel lightweight architecture that effectively captures global and local features with lower computational cost than existing transformer-based models.

Findings

Outperforms CNN and transformer-based methods in DSC and HD metrics

Achieves better segmentation of complex and small structures

Offers smaller size, lower computation, and faster training and inference

Abstract

Medical image segmentation is crucial for disease diagnosis and monitoring. Though effective, the current segmentation networks such as UNet struggle with capturing long-range features. More accurate models such as TransUNet, Swin-UNet, and CS-UNet have higher computation complexity. To address this problem, we propose GCtx-UNet, a lightweight segmentation architecture that can capture global and local image features with accuracy better or comparable to the state-of-the-art approaches. GCtx-UNet uses vision transformer that leverages global context self-attention modules joined with local self-attention to model long and short range spatial dependencies. GCtx-UNet is evaluated on the Synapse multi-organ abdominal CT dataset, the ACDC cardiac MRI dataset, and several polyp segmentation datasets. In terms of Dice Similarity Coefficient (DSC) and Hausdorff Distance (HD) metrics, GCtx-UNet outperformed CNN-based and Transformer-based approaches, with notable gains in the segmentation of complex and small anatomical structures. Moreover, GCtx-UNet is much more efficient than the state-of-the-art approaches with smaller model size, lower computation workload, and faster training and inference speed, making it a practical choice for clinical applications.