Implementation of Speech Recognition Based Eye Vision Test

Abstract

In today’s world, many underdeveloped countries often neglect eye care due to lack of awareness. Our idea describes a speech recognition-based eye test which is capable of performing the entire Snellen eye test on its own and estimates one's visual acuity. The test uses a few random letters like "A", "E", "I", "R", and "L" with the same 6 possible text sizes for each of these letters. The user speaks his/her guess into the microphone. The speech recognition portion uses energy thresholding. The system will then determine the next step based on whether the user guessed the displayed letter correctly or not. If the user guesses enough correctly, the text size will continue getting smaller and If the user guesses too many letters incorrectly, the system will display the result corresponding to the current text size. The thresholds for amount guessed and amount guessed correctly are set in the code. The software used for this Project is Python for Raspberry Pi