GPU-Net: Lightweight U-Net with more diverse features
This work addresses the need for efficient segmentation models in medical imaging, offering a plug-and-play module for existing methods, though it is incremental as it builds on U-Net.
The authors tackled the problem of efficient medical image segmentation by proposing GPU-Net, a lightweight U-Net variant that uses Ghost modules and ASPP to learn more diverse features, achieving better performance with over 4 times fewer parameters and 2 times fewer FLOPs.
Image segmentation is an important task in the medical image field and many convolutional neural networks (CNNs) based methods have been proposed, among which U-Net and its variants show promising performance. In this paper, we propose GP-module and GPU-Net based on U-Net, which can learn more diverse features by introducing Ghost module and atrous spatial pyramid pooling (ASPP). Our method achieves better performance with more than 4 times fewer parameters and 2 times fewer FLOPs, which provides a new potential direction for future research. Our plug-and-play module can also be applied to existing segmentation methods to further improve their performance.