An open-source food image embedding model, which is trained and can be loaded with caffe: http://caffe.berkeleyvision.org/

To execute the model, please copy the GRN layer implementation from this repo (.cpp, .cu and .h files) to the corresponding caffe directories and rebuild caffe as documented in http://caffe.berkeleyvision.org/installation.html.

Model download: https://s3.amazonaws.com/cornell-tech-sdl-fooddist/siamese_vgg_food_train_iter_100000.caffemodel

For open-source Yum-me system, please refer to: https://github.com/ylongqi/yumme

Contact: Longqi Yang, [email protected]

Reference:

Longqi Yang, Cheng-Kang Hsieh, Hongjian Yang, JP Pollak, Nicola Dell, Serge Belongie, Curtis Cole, Deborah Estrin, "Yum-me: A Personalized Nutrient-based Meal Recommender System" ACM Transactions on Information Systems (TOIS), 36.1 (2017): 7.