Hand Gesture Based Control Interface for Smartphones

EasyChair Preprint 7114

Abstract

This work provides various description of different methods of improving hand gesture recognition algorithms. A vision-based Hand Gesture Recognition system had been useful to recognize hand gesture in air. The work realizes the segmentation of hand gestures by establishing the skin color model and AdaBoost classifier based on CNN according to the particularity of skin color for hand gestures, as well as the denaturation of hand gestures with one frame of video being cut for analysis. In this regard, the human hand is segmentd from the complicated background, the real-time hand gesture tracking is also realized by CamShift algorithm.  To extract the features of air gesture we used statistical technique which is Principal Component Analysis and Machine Learning Algorithms such as Convolutional Neural Network and Adaboost for Object and Motion Detections. The idea here is to recognize gestures without the need to connect to a computer in which a database is located to perform training process. With this system, all steps can be done with optimum accuracy and the results were much improved. The work comprised of image acquisition, image processing, and after application of required machine learning algorithms our accuracy was close to 88.26%.

Keyphrases: CNN, Hand Gesture, control interface

@booklet{EasyChair:7114,
  author    = {Shubham Pandey and Abir Karmakar},
  title     = {Hand Gesture Based Control Interface for Smartphones},
  howpublished = {EasyChair Preprint 7114},
  year      = {EasyChair, 2021}}