Abstract
With the exponential growth of the presence of sport in the media, the need for effective classification of sports images has become crucial. The traditional approaches require carefully hand-crafted features, which make them impractical for massive-scale data and less accurate in distinguishing images that are very similar in appearance. As the deep learning methods can automatically extract deep representation of training data and have achieved impressive performance in image classification, our goal was to apply them to automatic classification of very similar sports disciplines. For this purpose, we developed a CNN-TL-DE method for image classification using the fine-tuning of transfer learning for training a convolutional neural network model with the use of hyper-parameter optimization based on differential evolution. Through the automatic optimization of neural network topology and essential training parameters, we significantly improved the classification performance evaluated on a dataset composed from images of four similar sports—American football, rugby, soccer, and field hockey. The analysis of interpretable representation of the trained model additionally revealed interesting insights into how our model perceives images which contributed to a greater confidence in the model prediction. The performed experiments showed our proposed method to be a very competitive image classification method for distinguishing very similar sports and sport situations.
Highlights
A huge increase of the amount of non-textual information available in electronic form in mass media requires dealing with it as a major source of content [1]
To solve the problem of classifying similar sports images into proper disciplines, we developed the adaptive convolutional neural networks (CNNs) training method, which is based on an adopted CNN architecture, trained using the transfer learning approach where the hyper-parameters for fine tuning are optimized with the use of differential evolution (DE)—we named the method CNN-TL-DE
CNN-TL-DE method with the CNN method for training the CNN model in an conventional manner and CNN-TL method that performs transfer learning upon the same CNN architecture
Summary
A huge increase of the amount of non-textual information available in electronic form in mass media requires dealing with it as a major source of content [1]. As image classification from the ML algorithms point of view is a demanding and complex task, researchers tried to improve the performance of traditional neural networks by introducing more hidden levels with multi layer perceptrons [20] or creating ensembles of neural networks [21]. Significant progress in this area has been achieved by the introduction of CNNs [22], which have become the de facto standard for classifying images with the improvement of computing power and the proliferation of large amounts of image data on which the advanced. The equivariance of convolutional layers enables memorizing the position of a certain feature in the input images, which can be recognized even when their position change
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
