Talk by Prof. Atsuto Maki on Dense FixMatch: a simple semi-supervised learning method for pixel-wise prediction tasks
Dense FixMatch: a simple semi-supervised learning method for pixel-wise prediction tasks
We discuss Dense FixMatch, a simple method for online semi-supervised learning of dense and structured prediction tasks combining pseudo-labeling and consistency regularization via strong data augmentation. It is an application of FixMatch enabled beyond image classification by adding a matching operation on the pseudo-labels. This allows us to still use the full strength of data augmentation pipelines, including geometric transformations.
We evaluated it on semi-supervised semantic segmentation on Cityscapes and Pascal VOC with different percentages of labeled data, and ablated design choices and hyper-parameters. Dense FixMatch significantly improves results compared to supervised learning using only labeled data, approaching its performance with 1/4 of the labeled samples.
In the talk we will also brief some other related activities from our school if time permits.
Prof. Atsuto Maki, School of Electrical Engineering and Computer Science, KTH Royal Institute of Technology
16:00 - 17:00 / Wednesday, May 10th, 2023
Onsite(NII, National Institute of Informatics)
SUGIMOTO Akihiro - Digital Content and Media Sciences Research Division, Professor / Vice Director-General