AR Image Synthesis Using Cached Recognition to Prevent Flickering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional augmented reality techniques face issues with misalignment and flickering of virtual objects due to time-consuming marker recognition processes, leading to a lack of realism in synthesized images.
Innovation Solution
A computer-readable storage medium with an image display program that repeatedly recognizes and synthesizes virtual objects with captured images, using a process result from a successful recognition process even if the current recognition fails, to prevent flickering and misalignment, ensuring a more realistic display.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If marker recognition process is executed repeatedly to track object position, then object position can be updated, but recognition time increases causing CG image to lag behind captured image
Solution Approach 1:
The system performs preliminary actions by capturing multiple images at different times before the final recognition is needed. When recognition fails for the current image, the system uses the last successful recognition result from a previous image, thereby preparing advance the recognition outcome to avoid lag and flickering.
Solution Approach 2:
The system implements feedback by continuously monitoring recognition success and using the results to determine whether to update the CG image position. When recognition fails, the system feedbacks by maintaining the last valid recognition result, creating a stable reference that prevents display flickering and ensures continuous smooth display.
2Duration of action of moving object
If CG image is generated and synthesized with captured image in real-time, then display smoothness is improved, but misalignment occurs when camera moves and captured image changes
Solution Approach 1:
The system performs preliminary recognition and calculation of three-dimensional position before the camera moves to a new position. By preparing the recognition result in advance and using it for CG image generation, the system ensures that the CG image is synthesized at the correct position even when the camera is moving, preventing misalignment while maintaining display continuity.
Solution Approach 2:
The system uses feedback from the recognition process to adjust the CG image position. When the camera moves and the captured image changes, the system feedbacks by updating the recognition result and recalculating the three-dimensional position, thereby maintaining accurate alignment between the CG image and the captured image throughout the movement.
3Adaptability or versatility
If recognition process fails intermittently, then system should handle failure gracefully, but CG image displays flickering when recognition fails
Solution Approach 1:
The system performs preliminary preparation by storing the last successful recognition result before a failure occurs. When recognition fails, this pre-stored result is used to continue generating the CG image, thereby preventing flickering and maintaining display stability while gracefully handling recognition failures.
Solution Approach 2:
The system implements feedback by monitoring recognition success and using the results to determine whether to update the CG image. When recognition fails, the system feedbacks by maintaining the last valid recognition result, creating a stable reference that prevents display flickering and ensures continuous smooth display.
4Measurement precision
If marker recognition is performed for each captured image, then position accuracy is maintained, but processing time increases causing time lag
Solution Approach 1:
The system performs preliminary recognition on captured images and stores the results for later use. By preparing recognition results in advance and reusing them when needed, the system maintains position measurement accuracy while reducing the processing time required for each individual image, thereby improving overall processing throughput.
Solution Approach 2:
The system uses feedback from successful recognitions to inform subsequent processing decisions. By caching and reusing valid recognition results, the system avoids redundant processing and maintains high accuracy while improving productivity through efficient use of previously obtained information.
Data Source
AI summary
A game apparatus obtains a captured image from a camera, and repeatedly executes a recognition process on the obtained captured image for recognizing a predetermined object included in the captured image. The game apparatus repeatedly generates and displays a synthesized age obtained by synthesizing an image of a virtual object generated by using a process result of the recognition process with a captured image which is a subject of the recognition process. Here, if the recognition process fails, a captured image which is a subject of a recognition process which has succeeded before is synthesized with an image of a virtual object generated by using a process result of the recognition process.