Implementation of Real Time Static Hand Gestures Recognition for Sign Language

Abstract

The paper propose that uses the computer vision of hand recognition. The camera records live video streams, where the image is taken with the help of the interface. The system is train for every type of hand gesture (one, two, three, four, and five) at least once. After that a test action is given and the system tries to detect it. A proposed system in which hand gestures are detected using image processing. The system detects the number of fingers. The system finds separate fingers above the palm. The system first detects skin color in an image using a filter. The image goes through various steps of image editing to give the right amount of fingers. The system detects the nearest point from the decision point [1]. The system deletes the image according to the centroid point. After that many steps are applied to adjust the image to the effective image so that the fingers are exposed. The system finally detects the number of fingers and displays the calculation to the user. Research has been done on many algorithms that can better distinguish hand gestures. It was found that the diagonal sum algorithm provided the highest accuracy measure. In the progression phase, an automated algorithm removes the background for each training action. The image is then converted into a binary image and taken statistics for all the separated elements of the image. This amount helps us to distinguish and distinguish different hand gestures.