Clothing-Similarity-using-Deep-Learning

Team members:

Aniket Kale (ankale@iu.edu), Ajinkya Pawale (ajpawale@iu.edu) ,Shivani Vogiral (svogiral@iu.edu)

Abstract:

• The focus of the project is to find the most suitable clothing matches for the current selection and choice of clothing using existing pre-trained Deep Learning based feature extractors.
• In addition, we also aim to use Siamese Network to find the similarity between images across different categories.
• The Siamese Network incorporated with the feature maps extracted from VGG16 was able to fully capture the visual similarity and clothing category matching the real intentions of the user.

Future implications of this project

• The similarity between clothing apparel can be used to aid clothing recommender systems.
• This is especially useful when there is less data available for the recommendation system engine to draw conclusions out of. And also aid in displaying similar patterns of apparel to the end user.