Neurosymbolic AI-based Framework For Sports Ball Identification Concerning Toddlers
Abhilasha Mangal, Tarjni Vyas, Rajesh Gupta, Sudeep Tanwar, Hossein Shahinzadeh · 2024
In today’s digital world, a vast amount of unstructured data is generated, primarily consisting of images and videos. Extracting meaningful information from these visuals is crucial for effective retrieval. Among the various fields where these images play a role, sports are essential in everyone’s life. This article introduces a system aimed at helping toddlers effortlessly understand sports balls. The proposed framework establishes a classification system for sports images, considering factors such as ball size, color, direction, dimension, and other vital details. Our approach uses object detection and image processing techniques to classify various sports balls, including cricket balls, basketball, volleyball, tennis balls, and football. To determine their positions and provide relevant information, we implemented the neuro-symbolic ai framework (NSAI) to provide symbolic reasoning question-answer capabilities to classify balls.