June 10, 2020
Prof. Gunhee Kim's team (Postdoctoral Researcher Junhyug Noh, Researcher Wonho Bae and Researcher Jinhwan Seo) participated in the Learning from Imperfect Data (LID) competition held at the CVPR 2020 with the support of Baidu.
In the challenge, we won the Weakly-supervised Object Localization (WSOL) section of the picture given, and the runner-up in the Weakly-supervised Semantic Segmentation (WSSS) section, winning 2,000 USD.
Weakness learning aims to achieve the level of supervised learning given full information even if only a part of the label is given in the learning process.
Further details can be found at https://lidchallenge.github.io.
Relating slides can be found at Slide.