Article: Artificial intelligence-based automatic identification and classification of diverse sports using advanced deep learning models Journal: International Journal of Information and Communication Technology (IJICT) 2025 Vol.26 No.23 pp.91 - 113 Abstract: The study examines state-of-the-art artificial intelligence (AI) methodologies aimed at developing sports image classification as it affects multimedia management as well as recommendation algorithms and sport data analysis capabilities. The sports industry is witnessing unprecedented growth, fuelled by advancements in technology, and the exponential rise of digital content. The vast quantity of sports-related media requires critical management for improved accessibility for user engagement capabilities. AI brings transformative automation capabilities through its ability to tackle these sorts of tasks. Deep learning applications show outstanding performance for resolving intricate classification challenges. This research developed a sports image classification framework using deep neural networks (DNNs) and analysed two pre-trained models ResNet-50 and MobileNet for performance comparisons. The DNN model demonstrated outstanding performance metrics through 98% accuracy which matched its precision and recall and F1-scores. DNN proved the most suitable solution when compared to pre-trained models ResNet-50 and MobileNet. Inderscience Publishers - linking academia, business and industry through research

Title: Artificial intelligence-based automatic identification and classification of diverse sports using advanced deep learning models

Authors: Yuan Zheng; Long Cai

Addresses: Weinan Normal University, Shaanxi, Weinan, 714000, China ' Weinan Normal University, Shaanxi, Weinan, 714000, China

Abstract: The study examines state-of-the-art artificial intelligence (AI) methodologies aimed at developing sports image classification as it affects multimedia management as well as recommendation algorithms and sport data analysis capabilities. The sports industry is witnessing unprecedented growth, fuelled by advancements in technology, and the exponential rise of digital content. The vast quantity of sports-related media requires critical management for improved accessibility for user engagement capabilities. AI brings transformative automation capabilities through its ability to tackle these sorts of tasks. Deep learning applications show outstanding performance for resolving intricate classification challenges. This research developed a sports image classification framework using deep neural networks (DNNs) and analysed two pre-trained models ResNet-50 and MobileNet for performance comparisons. The DNN model demonstrated outstanding performance metrics through 98% accuracy which matched its precision and recall and F1-scores. DNN proved the most suitable solution when compared to pre-trained models ResNet-50 and MobileNet.

Keywords: artificial intelligence; sports classification; game; deep neural network; DNN; feature extraction.

DOI: 10.1504/IJICT.2025.147123

International Journal of Information and Communication Technology, 2025 Vol.26 No.23, pp.91 - 113

Received: 12 Apr 2025
Accepted: 13 May 2025
Published online: 10 Jul 2025 *

Title: Artificial intelligence-based automatic identification and classification of diverse sports using advanced deep learning models

Keep up-to-date