Implementation and evaluation of DWT and MFCC based ISL gesture recognition
Neha Singh, Neha Baranwal, G. C. Nandi · 2014
In recent days, Indian Sign Language (ISL) has been assumed to have more appealing gestures for speech and hearing impaired community present in India. It helps us to understand the inherent meaning of this language for establishing a gesture based communicating system. In this paper a novel Discrete Wavelet Transform (DWT) and Mel Frequency Ceptral Coefficients (MFCC) based ISL gesture recognition technique is proposed. Here DWT is used for reducing the dimensionality of the data as well as for finding the most appropriate boundary points of hand gestures. After that MFCC is applied as a feature extraction technique for finding the spectral envelope of each image frame. This spectral envelope quality is useful for recognizing hand gestures in complex environment. Here support vector machine (SVM) and K Nearest Neighbour (KNN) is used as a classifier for classifying an unknown gesture. From experimental results it has been observed that DWT with MFCC provides high recognition rate with SVM as compare to KNN.