CSC-Unet: A Novel Convolutional Sparse Coding Strategy Based Neural Network for Semantic Segmentation

Abstract

It is still a challenging task to perform the semantic segmentation with high accuracy due to the complexity of real picture scenes. Many semantic segmentation methods based on traditional deep learning insufficiently captured the semantic and appearance information of images, which put limit on their generality and robustness for various application scenes. Thus, in this paper, we proposed a novel strategy that reformulated the popularly used convolution operation to multi-layer convolutional sparse coding block in semantic segmentation method to ease the aforementioned deficiency. To prove the effectiveness of our idea, we chose the widely used U-Net model for the demonstration purpose, and we designed CSC-Unet model series based on U-Net. Through extensive analysis and experiments, we provided credible evidence showing that the multi-layer convolutional sparse coding block enables semantic segmentation model to converge faster, extract finer semantic and appearance information of images, and improve the ability to recover spatial detail information. The best CSC-Unet model significantly outperforms the results of the original U-Net on three public datasets with different scenarios, i.e., 87.14% vs. 84.71% on DeepCrack dataset, 68.91% vs. 67.09% on Nuclei dataset, and 53.68% vs. 48.82% on CamVid dataset, respectively. In addition, the proposed strategy could be possibly used to significantly improve segmentation performance of any semantic segmentation model that involves convolution operations and the corresponding code is available at https://github.com/NZWANG/CSC-Unet.

Keywords

• U-Net
• semantic segmentation
• deep learning
• convolution operation
• convolutional sparse coding (CSC)