International Journal of Engineering Research in Computer Science and Engineering (IJERCSE)

Comparison of CNN and Contour Algorithm for Number Identification Using Hand Gesture Recognition

Author : Adarsh Pathak ¹ Faraz Ahmed Khan ²

Date of Publication :17th May 2021

Abstract: This paper compares the performance of two methods for hand gesture recognition for number identification. The image is captured employing a web camera system and undergoes many process stages before recognition of the numbers. Some of these stages include capturing the images, noise elimination, application of the CNN, and contour algorithm to predict the number. Once the hand is been placed in the region of interest the CNN algorithm predicts the number and gives output in the frame using deep learning techniques, whereas the contour algorithm creates the boundary of the hand and predict the number using the convexity hull defects algorithm and gives the output in the frame. The proposed methods of CNN and Contour achieved the accuracy of 91.2% and 93.8% respectively.

Reference :

1. Deepak K. Ray, Mayank Soni, Prabhav Johri, Abhishek Gupta, “Hand Gesture Recognition using Python”, International Journal on Future Revolution in Computer Science & Communication Engineering ISSN: 2454-4248 Volume: 4 Issue: 6.
2. S.G.Rayo, “Hand Gestures and Hand Movement Recognition for Multimedia Player Control”, February 2008, unpublished thesis.
3. T.-N. Nguyen, D.-H. Vo, H.-H. Huynh, and J. Meunier, “Geometry-based static hand gesture recognition using support vector machine,” In Proc. The 13th International Conference on Control Automation Robotics & Vision, Singapore, December 10–12, 2014.
4. Hung-Yuan Chung, Yao-Liang Chung, Wei-Feng Tsai, “An Efficient Hand Gesture Recognition System Based on Deep CNN,” 2019 IEEE International Conference on Industrial Technology (ICIT).
5. T. Koizumi, M. Mori, S. Taniguchi, and M. Maruya, “Recurrent neural networks for phoneme recognition,” In Proc. The Fourth International Conference on Spoken Language Processing, PA, USA, October 3–6, 1996.
6. Chigozie Enyinna Nwankpa, Winifred Ijomah, Anthony Gachagan, and Stephen Marshall, “Activation Functions: Comparison of Trends in Practice and Research for Deep Learning”, arXiv:1811.03378v1 [cs.LG] 8 Nov 2018.
7. Guifang Lin, Wei Shen, “Research on convolutional neural network based on improved Relu”, 8th International Congress of Information and Communication Technology.
8. https://www.edureka.co/blog/convolutional-neuralnetwork/ (Date of access: 23rd March 2021)
9. Yanan Xu1, Dong-Won Park and GouChol Pok “Hand Gesture Recognition Based on Convex Defect Detection”, International Journal of Applied Engineering Research ISSN 0973-4562 Volume 12, Number 18 (2017) pp. 7075-7079.