Segmentation Algorithm to Determine Group for Hand Gesture Recognition

Abstract

The main principle of hand gesture is recognizing any forms of gesture in the form of alphabet letters. The goal is to help the disabled to communicate with each other. Our system runs in real time without the help of sensors, gloves, etc. With such lighting conditions, different conditions of human hand and background of shooting become a problem in the completion of the process. This research proposed a segmentation method to resolve these problems. The method begins with capturing a picture using a webcam, which is followed by the segmentation process. We also proposed several conditions of skin detection. In this research, the segmented image undergoes the extraction process, which adopts three forms of feature extraction, namely slimness, roundness, and rectangularity. The final step of the method is measuring the resemblance of the images data features using adaptive neuro fuzzy inference system.