Review of Chinese Sign Language Recognition Methods and Technologies
Chinese Sign Language has its own unique cultural connotations and complex expressions,and it is an important means for more than 30 million hearing-impaired people to integrate into society.Sign language recognition technology can assist individuals with hearing impairments in bridging communication gaps and establishing effective communication with those who can hear.Chinese sign language recognition methods have gone through roughly two stages of traditional technology recognition and modern intelligent recognition.The former mainly includes four stages of data collection,preprocessing,feature extraction,and classification recognition.Hidden Markov Models(HMMs),Support Vector Machines(SVM),and Dynamic Time Warping(DTW)are mainstream technologies.It achieves recognition based on hand data without relying on extensive sample data.The latter mainly combines deep neural network and artificial intelligence technology,emphasizing deep learning,transfer learning,and technology integration.The model is highly dependent on the amount of sample data.China has started to extensively develop various sign language corpora,but it requires further standardization and promotion.
sign language recognition technologiescorpus,deep neural network,transfer learning
万方数据
维普
