Outfit Recognition and Classification System Using Deep Learning

There are many cultures and fashion prevailing in the fashion world. Each one has their own outfit. We may have seen some of them in images but may not know what that outfit is. By classifying the outfit worn by the person we can get to know about the dress and also about the person. The accessories worn by the person can also be detected which will help us to know what they are and can help to investigate more about them. To detect them and to overcome that we propose a system which will detect the outfit based on the classified data and detect the type of outfit worn by the user. By classifying the type of outfits, a wide range of applications can be developed in the real time world. Different transfer learning methods like ResNet-50, ResNet-101, VGG-16 and Efficient Net has been examined along with CNN architecture and found out that Efficient Net architecture is highly accurate with 99% result.