A Novel Technique for Speech Recognition and Visualization Based Mobile Application to Support Two-Way Communication between Deaf-Mute and Normal Peoples

Table 4

Comparison of proposed methodology with state-of-the-art ASR systems.

ASR Systems	Methodology	Accuracy
Proposed Methodology ()	MFCC + HTK (8-state HMM)	97.9%
MSIAC (Liu et al., 2017) [15]	MFCC + GMM	Experiment 1	89.50%
		Experiment 2	85.92%
TAMEEM V1.0 (Abushariah, 2017) [45]	MFCC + Sphinx 3	92.36%
Speaker Identification System (Leu and Lin, 2017) [46]	MFCC + GMM	Experiment 1	94.08%
		Experiment 2	84.88%
Telugu Speech Signals (Mannepalli et al., 2016) [47]	MFCC − GMM	92%
AMAZIGH LANGUAGE (Elouahabi et al., 2016) [48]	MFCC + HTK (6-state HMM)	80%