Speech Recognition for Control of Optotype Characters of the Snellen Chart using LogMAR Transformation

The study was demonstrated a speech recognition engine would interpreted characters of the Snellen chart into a series of acoustic vectors. These vectors were analyzed using Discrete Fourier Transform (DFT) and a bank filter to eliminate human voice dynamics. The equivalent acoustic vector in the phonic dictionary had been compared and provided a result measurement as visual acuity. There were ten respondents who participated with voice testing. The visual acuity results were then compared with result taken by an Ophthalmologist with Snellen chart. The LogMAR transform was used to aid with measurement of visual acuity and determined exact measurement based on its principle. The Hidden Markov Model was used as the base algorithm of speech recognition technology since most continuous speech system in present-day was established using this model.