Augmentation Overlay Restriction via Feature Pattern Recognition
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In augmented reality applications, there is a need to restrict the overlay of virtual content on specific real-world objects or areas to prevent unwanted augmentation, such as advertisements or logos, while maintaining the object's appearance to the human eye and allowing for recognition through computer vision techniques.
Innovation Solution
Embedding predefined feature patterns into real-world objects or media that can be recognized by computing devices using computer vision techniques, allowing the system to restrict the overlay of augmentations on specific areas or objects by evaluating feature space representations of images captured in augmented reality mode.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If virtual content is overlaid on real-world objects in augmented reality, then information delivery and user engagement are improved, but unwanted augmentation on specific objects like advertisements and logos cannot be prevented
Solution Approach 1:
The system performs preliminary actions by embedding feature patterns into real-world objects beforehand and pre-defining restriction zones associated with these patterns. When augmented reality content is to be displayed, the system has already prepared the knowledge base with pattern-restriction zone mappings, enabling immediate recognition and application of augmentation restrictions without real-time decision delays.
Solution Approach 2:
The system introduces an intermediary mechanism - the feature pattern recognition system - that acts as a mediator between the augmented reality content delivery system and the real-world environment. This intermediary evaluates images for feature patterns, determines restriction zones, and controls whether augmentations should be displayed, thereby preventing unwanted overlays without directly modifying the augmentation content itself.
2Measurement precision
If feature patterns are embedded into real-world objects to enable computer vision recognition, then precise control over augmentation placement is achieved, but the object's appearance may be altered
Solution Approach 1:
The system applies local quality by embedding feature patterns only in specific, localized regions of real-world objects rather than modifying the entire object. These patterns are placed in discrete locations where they can be detected by computer vision without significantly altering the overall appearance or essential characteristics of the object to the human eye.
Solution Approach 2:
The system utilizes parameter changes by adjusting the visual parameters of feature patterns to be imperceptible or minimally perceptible to humans while remaining detectable by computer vision algorithms. The patterns are designed with specific parameter ranges (size, contrast, complexity) that optimize machine recognition while maintaining natural appearance for human observers.
3Adaptability or versatility
If augmentation overlay is allowed on all areas, then maximum flexibility in content delivery is maintained, but control over placement on specific objects is lost
Solution Approach 1:
The system implements dynamics by making the augmentation display rules adaptive rather than static. The system dynamically evaluates each image for the presence of feature patterns and their associated restriction zones, and adjusts augmentation display decisions in real-time based on this evaluation. This allows the system to maintain flexibility for areas without restrictions while automatically applying controls where needed.
Solution Approach 2:
The system employs feedback mechanisms by continuously monitoring the visual environment for feature patterns and using this information to adjust augmentation placement decisions. The feedback loop involves capturing images, evaluating them for feature patterns, determining restriction zones based on pattern recognition, and then using this information to control whether augmentations should be overlaid in specific areas.
Data Source
AI summary
Methods and apparatuses are provided for restricting overlay of an augmentation. A method may include evaluating a feature space representation of an image for presence of a predefined feature pattern. The method may further include recognizing, based at least in part on the evaluating, that the predefined feature pattern is present in the feature space representation of the image. The method may additionally include, in response to recognizing that the predefined feature pattern is present in the feature space representation of the image, restricting overlay of an augmentation on a portion of the image corresponding to the recognized predefined feature pattern. Corresponding apparatuses are also provided.