A Hierarchical Approach for Spanish Sign Language Recognition: From Weak Classification to Robust Recognition System

Approximately 5% of the world's population has hearing impairments and this number is expected to grow in the coming years due to demographic aging and the amount of noise we are exposed to. A significant fraction of this population has to endure severe impairments even since their childhood and sign languages are an effective mean of overcoming this barrier. Although sign languages are quite widespread among the deaf community, there are still situations in which the interaction with hearing people is difficult. This paper presents the sign language recognition module from an ongoing effort to develop a real-time Spanish sign language recognition system that could also work as a tutor. The proposed approach focuses on the definitions of the signs, first performing the classification of their constituents to end up recognizing full signs. Although the performance of the classification of the constituents can be quite weak, good user-independent sign recognition results are obtained.