Hand Gesture Recognition Using Otsu's Method

In this smart generation, there are various means for communicating, such as people communicate through phones via voice and text. The same way we can communicate with electrical devices through Arduino and with the help of computer. This work presents a method for a human PC interface through hand motion acknowledgment that can perceive static signals. The target of this postulation is to build up a calculation for acknowledgment of hand motions with sensible precision. Experimental setup of system uses a laptop camera which is fixed at certain position and snapshots are taken in red, blue, and green (RGB) color space. The input from the camera undergoes pre processing, segmentation, feature extraction, pixel shifting and classification process. The captured RGB image is converted into gray-scale image and background noise is ignored by using the plain background. The image is segmented using Otsu's segmentation method. The features are extracted using SURF function. The images are shifted towards reference end and are subtracted. This framework can dependably perceive single-hand motions continuously and can accomplish estimated 75% recognition rate in plain background with a “minimal possible constraints” approach.