<div dir="ltr"><b>Learning from Limited and Imperfect Data (L2ID) Challenges</b><br>In conjunction with the L2ID Workshop at CVPR 2021<br>June 20, 2021 (Full Day, Virtual Online)<br><div><a href="https://l2id.github.io/" target="_blank">https://l2id.github.io/</a></div><br>******************************<br><b>CALL FOR PARTICIPATION</b><br>As
part of the CVPR 2021 L2ID workshop, we have two sets of challenges on
weakly supervised learning and multi-domain few-shot classification.<br>We invite you to participate! Winners will be able to present their works during the workshop.<br><br><b>Localization</b><br><div>Check the rules at <a href="https://l2id.github.io/challenge_localization.html" target="_blank">https://l2id.github.io/challenge_localization.html</a></div><br>Track 1 - Weakly Supervised Semantic Segmentation<br>This track targets on learning to perform object semantic segmentation using image-level annotations as supervision.<br><br>Track 2 - Weakly supervised product detection and retrieval<br>Given
a photo containing multiple product instances and a user-provided
description, the track aims to detect the boxes of each product and
retrieve the correct single product image in the gallery.<br><br>Track 3 - Weakly-supervised Object Localization<br>This track targets on making the classification networks be equipped with the ability of object localization.<br><br>Track 4 - High-resolution Human Parsing<br>This track aims to recognize human parts within high-resolution images by learning with low-resolution ones.<br><br><b>Multi-Domain Few Shot Classification</b><br>Check the rules at <a href="https://l2id.github.io/challenge_classification.html" target="_blank">https://l2id.github.io/challenge_classification.html</a><br><br>Track 1 - Cross Domain, small scale<br>This
track calls for the development of cross-domain few-shot learning
models starting from multiple sources and with no explicit label overlap
between sources and target.<br><br>Track 2 - Cross Domain, large scale<br>In
this track additional datasets to both source and target datasets have
been added for participants with sufficient compute resources. In
addition to the multiple sources, *multiple tasks* from which to draw
source data or models are provided.<br><br>Track 3 - Cross Domain, larger number of classes<br>In this track, the “wayness” of few-shot learning is increased, bringing it closer to semi-supervised learning.<br><br>******************************<br><b>IMPORTANT DATES</b><br><br>Submission Deadline: May 14th<br>Leaderboard Published / Invitations Sent: May 21th<br><br>******************************<br><b>WORKSHOP ORGANIZERS:</b><br>Zsolt Kira (Georgia Tech, USA)<br>Shuai (Kyle) Zheng (Dawnlight Technologies Inc, USA)<br>Noel C. F. Codella (Microsoft, USA)<br>Yunchao Wei (University of Technology Sydney, AU)<br>Tatiana Tommasi (Politecnico di Torino, IT)<br>Ming-Ming Cheng (Nankai University, CN)<br>Judy Hoffman (Georgia Tech, USA)<br>Antonio Torralba (MIT, USA)<br>Xiaojuan Qi (University of Hong Kong, HK)<br>Sadeep Jayasumana (Google, USA)<br>Hang Zhao (MIT, USA)<br>Liwei Wang (Chinese University of Hong Kong, HK)<br>Yunhui Guo (UC Berkeley/ICSI, USA)<br>Lin-Zhuo Chen (Nankai University, CN)</div>