Utilizing a Novel Convolutional Neural Network (CNN) Algorithm instead of a Genetic Algorithm (GA) based CNN Algorithm to Improve Accuracy, a Smart System to Recognize Human Actions.

Abstract

Aim: The major purpose of the research is to recognize human actions using the Convolutional Neural Network algorithm(CNNT) in comparison with Genetic algorithm(GT) for the TRAIN dataset. Materials and Methods: Recognition of actions in humans is recognized using Novel Convolutional Neural Network algorithm (N=20) and Genetic algorithm (N=20). Novel Convolutional Neural Network algorithm is a supervised machine learning, Deep learning recognition technique, it is basically needed for categorization and recognition because of its high accuracy. Genetic algorithm is a Heuristic search algorithm used to solve search and optimization. Human acts identification dataset  is used for recognizing human acts. Results: The precision of human acts identification using the NCNNT is 99.9% and Genetic Technique(GT) is 97.9%. There's a huge distinction between Novel Convolutional Neural Network Technique(NCNT) and Genetic algorithm with 0.003(p<0.001). Conclusion: Novel CNNT seems more accurate than Genetic  Techniques(GT) in recognition of  human actions.