Wearable AR Object Selection via Pre-emptive Database Search
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current augmented-reality devices struggle to accurately and efficiently recognize and confirm object selections in real-time, leading to potential inaccuracies in subsequent searches or actions.
Innovation Solution
A wearable computing device and server system that records video data to detect movement corresponding to object selection, initiates a search on an object-model database before movement completion, and periodically updates the search, overlaying the object with object-models from the database, allowing for real-time confirmation and improved accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the device waits for movement completion before initiating object recognition, then measurement accuracy is improved, but response time deteriorates
Solution Approach 1:
The system initiates object recognition and database search before the user completes the selection movement. The processor begins querying the object-model database using initial video data while the user is still performing the selection gesture, thereby overlapping the recognition process with the selection process and eliminating idle waiting time.
Solution Approach 2:
The object recognition process continues continuously throughout the movement rather than being paused until movement completion. The system periodically updates the search based on ongoing video data, maintaining continuous useful action from the moment selection begins through completion, thus eliminating gaps in the recognition process.
2Measurement precision
If the device performs frequent updates during movement, then object recognition accuracy is improved, but device complexity increases
Solution Approach 1:
The system performs object recognition updates at periodic intervals during the user's selection movement rather than continuously. The processor queries the object-model database at regular frames or time intervals, balancing the need for accurate recognition with the computational burden of constant processing.
Solution Approach 2:
The system performs more recognition updates than the minimum single post-movement recognition. By initiating recognition early and performing multiple partial updates during movement, the system achieves higher accuracy through cumulative evidence while distributing the computational load across multiple smaller processing tasks rather than one large task.
Data Source
AI summary
Methods and devices for initiating, updating, and displaying the results of a search of an object-model database are disclosed. In one embodiment, a method is disclosed that includes receiving video data recorded by a camera on a wearable computing device and, based on the video data, detecting a movement corresponding to a selection of an object. The method further includes, before the movement is complete, initiating a search on the object of an object-model database. The method still further includes, during the movement, periodically updating the search and causing the wearable computing device to overlay the object with object-models from the database corresponding to results of the search.