iMat Fashion 2021

Overview

As part of the FGVC8 workshop at CVPR 2021 we are conducting the iMaterialist - Fashion track 2021 apparel instance segmentation & fine-grained attribute classification competition. The goal of this competition is to push the state of the art of the novel fine-grained segmentation task by joining forces between the fashion and computer vision communities.

To capture the complex structure of fashion objects and ambiguity in descriptions obtained from crawling the web, our standardized taxonomy contains 46 apparel objects (27 main apparel items and 19 apparel parts), and 294 related fine-grained attributes. Secondly, a total of 50K clothing images (with both segmentation masks and fine-grained attributes) in daily-life, celebrity events, and online shopping are labeled by both domain experts and crowd workers for fine-grained segmentation.

In this competition, we challenge you to develop algorithms that will help with an important step towards automatic product detection – to accurately assign segmentations and attribute labels for fashion images. Individuals/Teams with top submissions will be invited, at the discretion of the workshop organizers, to present their work at the FGVC8 workshop.

Competition

Start Date - March 31st

End Date - May 26th

Kaggle URL - https://www.kaggle.com/c/imaterialist-fashion-2021-fgvc8/

Organizers

Menglin Jia, (Cornell University), Mengyun (David) Shi (Hearst Magazines), Mikhail Sirotenko (Google AI), R.J. Lehman (Hearst Magazines), Serge Belongie (Cornell University and Cornell Tech)

Acknowledgements

The iMat-Fashion Challenge 2021 is sponsored by Google AI, CVDF, and Fashionpedia.