Abstract: Continuous Sign Language Recognition (CSLR) aims to transcribe the signs of an untrimmed video into written words or glosses. The mainstream framework for CSLR consists of a spatial module ...