Content Based Image Retrieval for Fashion by using recent advances deep learning approaches

Abstract

Fashion Visual Search is to find other similar images in the product catalog for a consumer’s uploaded images based on visual appearance. In this report we propose using recent advanced deep learning methods and compare them with a ranking evaluation metric. The proposed feature extractors are trained and evaluated in two different datasets: the first contains many clothing images from both customer and shop and the other uses pure clothing items with a simple white background.