Application of Modern Sports Training using Logish Activation Function based Convolutional Neural Network

With respect to enhance the teaching quality as well as training effectiveness of the sports training, Human Computer Interaction (HCI) and Artificial Intelligence (AI) technologies is utilized to develop the expert system. An area of the sports is one of the significant fields of AI technology which have been developed. This research proposes the AI technology of Logish Activation Function based Convolutional Neural Network (LACNN) for the application and estimation of the modern sports training for college students. The logish activation function permits the network to emphasis more on the significant parts of an input data like key moments or movements in sports training videos, so it leads to enhance the model accuracy. The results shows that the proposed LACNN approach attains the accuracy of accuracy of 99.52%, precision of 97.82%, recall of 96.35% and Fl-score of 98.67% respectively. These obtained values show the superior outcome when compared to the existing methods like CNN and Video-based Effective Visualization Framework (VEVF).