A Study on Federated Learning-based Computer Vision Models for AR-VR Applications

Abstract

Federated learning has emerged as a promising approach to address privacy concerns and data decentralization in machine learning applications. This study investigates the application of Federated Learning in developing computer vision models for Augmented Reality (AR) and Virtual Reality (VR) applications. The research focuses on leveraging distributed data sources to train collaborative models without centralized data aggregation, thereby preserving user privacy and data security. The study uses Federated learning techniques to enhance AR-VR environments' object detection, recognition, and tracking capabilities. The research methodology involves the development of a federated learning framework tailored for computer vision tasks in AR-VR applications, considering the unique challenges posed by decentralized data sources. Evaluation metrics such as model accuracy, convergence speed, and communication efficiency will be analyzed to assess the performance of the federated learning-based computer vision models. The study aims to advance privacy-preserving machine learning techniques in AR-VR domains while maintaining high model accuracy and efficiency. The findings from this research are expected to provide valuable insights into the practical implementation of Federated Learning for enhancing computer vision capabilities in AR-VR applications.