Clothing Retrieval via Co-segmentation and Feature Matching

As the problem of unsatisfied accuracy of the clothing retrieval, a clothing retrieval method was presented via co-segmentation and feature matching. Firstly, the query image was segmented with an auxiliary dataset. Then, the color features and the bundled features of clothing regions were extracted after co-segmentation. The similarities of the features between the query image and the ones in retrieval dataset were calculated. At last, the retrieval results were output in descending order according to the similarity. Experimental results show that the proposed method achieves a high segmentation accuracy and retrieval accuracy. The segmentation accuracy is improved approximately 15% comparing with the known segmentation methods, and the retrieval accuracy is greatly improved in every clothing categories.