Lexicon-Free Fingerspelling Recognition from Video: Data, Models, and Signer Adaptation | Read Paper on Bytez