Method and System for recognizing continuous sign language based on computer vision

摘要

PURPOSE: A method and a system for recognizing successive sign language based on computer vision are provided to efficiently recognize had gesture and output the meaning of the sign language into voice signal. CONSTITUTION: A data obtaining part obtains color image data of a person talking with the hands(S510). A pre-processor processes hand video data for obtaining hand locus data and hand gesture data(S511-S513). A hand motion dividing part divides the hand locus data into individual sign language sentences and again divides into sign language(S514-S515). A hand motion classifying part classifies the least unit of sign language by using the individual sign language and the hand gesture data(S516-S517). A sign language interpreting part recognizes sign language words by combining the classified unit of sign language(S518). The sign language interpreting part interprets the sign language words into the sign language sentences by considering a sign language grammar(S519-S520). A sound generating part outputs the interpreted sign language sentences in a voice(S521).