Hand Gesture Recognition Using Convolutional Neural Network and Diagonal Sum Algorithm

  • Yasir Reyaz Mir, Ravinder Pal Singh, Monika Mehra

Abstract

The research presents a computer-based vision application to identify hand gestures. A camera records a live video stream that is utilised to build an interface snapshot. For every five varieties of count-hand movements, the methodology is taught at least once (one, two, three, four, and five). After that, the system attempts to recognise a test gesture. The Convolutional Neural Network was used since its algorithm is self-developed and during the pre-processing stage it removed the backdrop for every training move. The image is turned into a binary image and the sums of the diagonal components of the picture are computed. This amount helps to distinguish and classify distinct hand movements. Data gloves or markers were used in prior systems to enter data into the system. When it comes to utilising the system, I have no such constraints. The user may intuitively move his/her hand in front of the camera.

Published
2021-10-14
How to Cite
Yasir Reyaz Mir, Ravinder Pal Singh, Monika Mehra. (2021). Hand Gesture Recognition Using Convolutional Neural Network and Diagonal Sum Algorithm. Design Engineering, 4391-4398. Retrieved from http://www.thedesignengineering.com/index.php/DE/article/view/5399
Section
Articles