Immersive AR Supervision Boundary Adjustment With ML-Based Data Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing augmented reality (AR) systems face challenges in efficiently generating and maintaining immersive supervision boundaries due to the burden of transmitting, storing, and processing large image/video data, which strains computer networks, storage resources, and computational resources.
Innovation Solution
Implementing machine learning techniques that utilize predictive signals from location-based and sensory data to dynamically generate and maintain immersive AR supervision boundaries, reducing the need for large image/video data transmission, storage, and processing.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If large image/video data is transmitted and processed to generate supervision boundaries, then the accuracy and reliability of supervision boundaries is improved, but the burden on computer networks, storage resources, and computational resources increases
Solution Approach 1:
The patent extracts only the essential features and metadata from image/video data rather than processing the complete raw data. The system identifies and processes key elements such as object detection results, tracking data, and event markers, leaving out the bulk of redundant visual information. This extraction approach maintains supervision boundary reliability while significantly reducing network transmission burden, storage requirements, and computational processing loads.
Solution Approach 2:
The system performs preliminary processing of image/video data at the source device before transmission. Features such as object detection, tracking identifier assignment, and event detection are completed in advance, so that only processed results and essential metadata need to be transmitted to the server. This preliminary action reduces the data burden on networks and servers while maintaining the reliability of supervision boundaries through pre-validated feature extraction.
2Adaptability or versatility
If real-time monitoring data is continuously processed to dynamically adjust supervision boundaries, then the adaptability of the system is improved, but the computational resources required increase
Solution Approach 1:
The system implements periodic processing of monitoring data at strategic intervals rather than continuous real-time processing. Event-triggered updates occur based on specific conditions such as detected events, boundary violations, or significant changes in tracking data. This periodic action maintains system adaptability by responding to critical changes while significantly reducing computational energy consumption compared to continuous processing.
Solution Approach 2:
The system employs edge computing capabilities where the AR device performs local processing of monitoring data to generate and update supervision boundaries autonomously. The device uses on-device machine learning models to process tracking data and adjust boundaries without requiring constant server communication and processing. This self-service approach maintains adaptability through local real-time adjustments while minimizing the computational energy burden on centralized servers.
Data Source
AI summary
Various embodiments of the present invention provide methods, apparatuses, systems, computing devices, computing entities, and/or the like for performing AR processing using at least one of: (i) an exposure classification machine learning framework comprises a real-time embedding machine learning model, an exposure clustering machine learning model, and a cluster mapping machine learning model, (ii) supervision boundary reliability scores determined based at least in part on covered subsets for immersive AR supervision boundaries, (iii) supervision boundary reachability scores determined based at least in part on response node locations and least reachable locations of immersive AR supervision boundaries, or (iv) performing one or more AR interaction actions based at least in part on an updated immersive AR supervision boundary that is generated by reducing a supervision boundary area of the current immersive AR supervision boundary so that the supervision boundary reliability score satisfies the supervision boundary reliability score threshold.