Fetching the paper…

Cross-Domain First Person Audio-Visual Action Recognition through Relative Norm Alignment · Around