A research team improved the AI recognition accuracy of word-level sign language recognition by adding data such as the signer's hand and facial expressions, as well as skeletal information on the position of the hands relative to the body.
↧