2 datasets found

The LRW dataset is an English language lip reading dataset, containing 500 different words, each spoken by over 1,000 persons.

Speaker verification systems experience significant performance degradation when tasked with short-duration trial recordings. To address this challenge, a multi-scale feature...