A static hand gesture recognition system to recognize the total number of fingers

The main purpose of gesture recognition systems is to understand important expressions of motion by humans which involve hands, head, face, arms, or body. It is of major importance in designing an expert interface between humans and computers. We take into account the two possible fixed geometries to work on. The proposed methods follow the procedure of Preprocessing which helps with noise removal and image enhancement; Segmentation of hand region uses skin likelihood method to extract skin color; Feature extraction uses morphological and geometry based functions to extract the fingers; and active fingers are counted by method of Rule based Classification. In order to test performance, an experiment is conducted using standard and a self-generated dataset of images. The accuracy achieved on these dataset is greater than the similar state-of-the-art.