The implemented core codes of CSC-Unet are open here.

If you used our CSC-Unet codes, please cite our following papers:

Tang H, He S, Yang M, et al. (2024). CSC-Unet: A Novel Convolutional Sparse Coding Strategy based Neural Network for Semantic Segmentation. IEEE Access，doi：10.1109/ACCESS.2024.3373619.

Tang H, Shi J, Lu X, Yin Z, Huang L, Jia D, & Wang N. (2020, November). Comparison of Convolutional Sparse Coding Network and Convolutional Neural Network for Pavement Crack Classification: A Validation Study. In Journal of Physics: Conference Series (Vol. 1682, No. 1, p. 012016).

If you have any question or collaboration suggestion about our method, please contact [email protected].

The codes of various networks were tested in Pytorch 1.5 version or higher versions(a little bit different from 0.8 version in some functions) in Python 3.8 on Ubuntu machines (may need minor changes on Windows).

• 1. Clone this repo to local

git clone https://github.com/NZWANG/CSC-Unet.git

• 2. Download the experiment dataset from the link below, and put it into the directory: ./Datasets/CamVid/ ./Datasets/DeepCrack/ ./Datasets/Nuclei/
     1. CamVid: G. J. Brostow, J. Shotton, J. Fauqueur, and R. Cipolla, “Segmentation and recognition using structure from motion point clouds” http://mi.eng.cam.ac.uk/research/projects/VideoRec/CamVid/
     2. DeepCrack: Y. Liu, J. Yao, X. Lu, R. Xie, and L. Li, “DeepCrack: A deep hierarchical feature learning architecture for crack segmentation” https://github.com/yhlleo/DeepCrack
     3. Nuclei: https://www.kaggle.com/c/data-science-bowl-2018/overview
• 3. Set the hyper parameters of the experiment in cfg.py.
• 4. Run the code by command

python train.py