Page 105 - KBB_Sep_2020

User Groups
persons with Cerebral Palsy
Description and Functionality
SISOCA help in improving intelligibility of dysarthric speech by correcting articulatory errors of cerebral palsy speakers. The device acts as an interpreter that recognizes disordered speech and produces its intelligible equivalent. SISOCA includes (i) Assessment of the articulation errors made by the CP speakers and modeling errors, (ii) Speaker-dependent speech recognition system that converts unintelligible speech into text, (iii) Error correction system to correct the errors in the recognized text, with the assessment knowledge, using weighted finite state transducers (WFSTs), and a (iv) Text-to-speech synthesis system to synthesize the error corrected text into intelligible speech using a normal speaker’s voice.
State-of-the-art
SISOCA involves no keyboard or visual intervention and the technique adopted is language-independent. The tool developed as a portable mode of AASC device using Raspberry Pi works not only for isolated words but also for phrases.
Stage of Development
SISOCA involves no keyboard or visual intervention and the technique adopted is language-independent. The tool developed as a portable mode of AASC device using Raspberry Pi works not only for isolated words but also for phrases.
Specifications and Approximate Cost
SISOCA includes a Raspberry Pi, a collar microphone and a speaker. Dimensions: Raspberry Pi 4 – Height: 8.56 cm, Width: 5.65 cm, Depth 1.1 cm
Collar microphone: 17.6 x 16.8 x 4.2 cm; Speaker: variable size Estimated Cost: ₹5000/-
Technology Developer
Prof. P. Vijayalakshmi
SSN College of Engineering, Kalavakkam, Chennai-603110, Tamil Nadu vijayalakshmip@ssn.edu.in +91 9940415589
105