This paper focuses on unsupervised learning for lesion region segmentation in diabetic foot images, inspired by the human eye’s sensitivity to color differences in distinguishing lesions. A novel algorithm named Color Decoder (CDe) is proposed. The algorithm first segments the foot region based on skin color HSV thresholds, then uses a clustering method to differentiate significantly different color patches, and finally selects smaller regions among them as the target regions. This work is evaluated using the Diabetic Foot Ulcer Challenge 2024 datasets. The proposed method achieved a F1-Score of 0.2462 on validation set. When tested on the DFUC 2024 testing set, we achieved F1-Score of 0.2246. The code of this paper can be found on the following github link: https://github.com/HaoqianSong/DFUC2024/tree/master .