ASL CHARACTERS RECOGNITION USING PERSON SPECIFIC SUBSPACE METHOD

Abstract

This paper presents a person-specific subspace method for ASL characters recognition. To implement this person specific subspace method for ASL characters recognition the system first recognizes the faces and then uses skin-color segmentation method for hand poses segmentation. From the segmented hand poses the system recognizes ASL characters using both PCA and person-specific subspace method separately. Experiment result shows that the classification accuracy of person-specific subspace method is better than general PCA method. The experiment result also shows that person-specific subspace method is faster than general PCA method