Apparel Recommendation System using Transfer Learning

Chandrakala G. Raju, R. Dhruv Metha, P.A. Madhuri, Nidhi Jagadeesha and H.S. Ranjini Rao

Apparel classification involves classification ofclothes into various categories and suggesting similar clothes in that particular category. Apparel classification is very useful for people to identify and suggest the specific categories of clothing they want to purchase online. The proposed work has a model to optimize the process of searching for any kind of apparel we want by implementing image-based search. The classification is done based on the predefined set of classes such as Ladies Kurta, Saree, Pants, T-shirts and Footwear. Our image classification technique restricts the search to a subset of the entire dataset rather than an exhaustive search, i.e. the search is carried out in the small clusters that we have created rather than the entire dataset of images which in turn reduces the image search time. The proposed model has an average accuracy of 97.2%.

Volume 11 | 04-Special Issue

Pages: 1585-1592