As part of the FGVC4 workshop at CVPR 2017 we are also conducting the iMaterialist Challenge 2017, a fine-grained attribute recognition challenge for product images. With the growing popularity of online shopping, being able to recognize products from pictures is an important but challenging problem. The recognition is challenging because, for the same product, a picture can be taken under different lighting and viewing angle, various backgrounds, and different level of occlusion. One central piece for recognizing a product is to understand and obtain the attributes information of a product. In this competition, FGVC workshop organizers and Google invite you to join efforts in advancing technologies for fine-grained attribute recognition for products.


Data Released June 1, 2017

Submission Server Open June 1, 2017

Submission Deadline July 7, 2017

Winners Announced July 21, 2017

Competition Details

Checkout the iMaterialist Competition Github repo for the specifics of the dataset and download links. The final results will be evaluated using a top-5 accuracy metric. The competition itself is being hosted on Kaggle.


Thanks to the following people for their help organizing the dataset and competition: