Using gesture for interaction is a notion of natural human computer interaction. The use of gesture requires a correct identification of the intended pose by the system. Already numerous algorithms are working on it but not paying much attention on the complexity. Hence in this paper, a hybrid algorithm is being introduced to reduce the complexity of process also to make it insensitive with respect to the position of camera by making use of Gabor filer to get the structure of the hand followed by the Mel scaling in order to reduce the dimensions along with feature extraction using triangular filters. For the classification purpose Support vector machine has been used as a classifier. © 2012 IEEE.