This repo is the visualization of view normalization in "View-Normalized and Subject-Independent Skeleton Generation for Action Recognition". This work proposes a Variance Reduction (VaRe) framework to reduce the view- and subject-induced intraclass variances for skeleton-based action recognition.
This repo contains five videos. For each video, we illustrate the RGB video, the raw skeleton derived from diverse viewpoints, and the view-normalized skeleton generated by the proposed VN-GAN from the first column to the third column.
