Voiceless voice: An intelligent communication system for physically retarded communities

Speech is one of the common modes of communication and it is a process of transferring information from one entity to another but people who are paralyzed like amyotrophic lateral sclerosis (ALS) and disability like communication disorder are finding difficulties in communication with the external world. Studies show that Brain machine interface (BMI) provides a communication channel between the human brain and an external device. In this research work, a possible option for rehabilitation of such patients through brain machine interfaces which uses their active cognition capabilities to convey the imaginary thoughts in vocalized form has been proposed. EEG signals collected from 12 subjects during 19 mental tasks are used to test the classifiers. Spectral energy on the band power of the EEG signals is extracted for each frame. PCA features extracted from the spectral energy features are used the neural net to classify different mental tasks. Results obtained show average classification rates ranging from 93.15% to 94.74 %.