Skip to content

Deep Learning based visual product recommendation using CNN feature extraction and visual attention

Notifications You must be signed in to change notification settings

AaronCGoidel/icing

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

13 Commits
 
 
 
 
 
 

Repository files navigation

Icing

A ML based approach to suggest fashion products to users based on an input image.

Approach

An implementation of Complete the Look: Scene-based Complementary Product Recommendation (Kang 2019)

The goal is to predict the compatibility of a product image with a given scene image. Compatibility is measured both globally and locally.

Global compatibility is the overall compatibility of the product with the image. A feature vector is extracted from the scene image and the product using resnet. These are embedded using a small feed forward network. Global compatibility is the similarity between the two embeddings.

Local compatibility is the compatibility of the product with the different regions of the scene. The scene image is divided into subregions and a feature vector is extracted from each subregion. We then compute the attention weights of each subregion given the product category. Local compatibility is the weighted sum of the subregion embeddings.

About

Deep Learning based visual product recommendation using CNN feature extraction and visual attention

Resources

Stars

Watchers

Forks