Clothes Image Searching System Based on SIFT Features

One of the main difficulties in Clothes image searching is the variability of appearance of an object of interest. Clothes image feature extraction is a very active research area under this situation. Through build a Clothes concept semantic structure Clothes Tree Semantic Structure, the Clothes Image Searching System based on Scale Invariant Feature Transform(SIFT) is proposed in the paper. Experiment on an image database of about 5,000 general-purpose images, results show that our algorithm is more efficient than existing algorithms based on color histogram or corners features.