Human Action Recognition
M. BalajI
R. Kavin Kumar
B. Kishore
T. Saran Sujai
Dr.G. Singaravel
Keywords: OpenCV, CNN, Histogram.
Abstract
segmentation theory and a hand detection system implemented using Python with OpenCV. Hand gestures serve as a natural interface, which has motivated research in gesture taxonomies, representations, recognition methods/algorithms, and software platforms/frameworks. This project provides a detailed overview of these aspects. The growing acceptance and funding for multinational projects highlight the need for sign language solutions. Computer-based solutions are particularly important in today's technological age for the deaf community. Although researchers have been studying this problem for some time, promising results have emerged. This project presents a comprehensive review of vision-oriented sign recognition methodologies, emphasizing the importance of considering algorithm recognition accuracy for successful real-world implementation. The project matches a given image with dataset images containing numerous categories of sign gestures. A convolutional neural network (CNN) is implemented to improve accuracy. The project involves grayscale conversion, binary image conversion, histogram construction, and matching of the test image with dataset images. The coding language used is Python 3.7.
