1 paper
Salman Iqbal, Waheed Rehman
Zero-shot action recognition relies on transferring knowledge from vision-language models to unseen actions using semantic descriptions. While recent methods focus on temporal mode…