Improvement of Chinese Sign Language Translation System Based on Multi-Node Micro Inertial Measurement Unit

Keke Tu,Chong Pan,Jiye Zhang,Yufeng Jin,Jack Wang,Guangyi Shi
DOI: https://doi.org/10.1109/cyber.2015.7288216
2015-01-01
Abstract:This paper presents the improvement of Chinese sign language translation system based on studying in HMM algorithm, modeling and analysis. Based on the multi-node micro inertial measurement unit built by MEMS sensors, finger motion can be recoded and analyzed by computer. The computer preprocesses the data from the MEMS sensor nodes, including filtering noise removal and feature extraction. At last we train a classifier through HMM training process. Against 100 daily operations of sign language recognition experiments, the overall recognition rate is 90%. With the optimization and improvement of the algorithm, recognition accuracy and practicability will be greatly improved. We also designed a bidirectional translation system which can switch translation between Chinese sign language and voice freely.
What problem does this paper attempt to address?