Reference Image Slicing for AR Object Tracking
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Augmented reality systems face challenges in maintaining object recognition and tracking when the camera zooms in on an object, as insufficient reference features are detected, leading to loss of recognition and tracking capabilities.
Innovation Solution
The method involves generating reference features sets from slices of the original reference image, allowing for effective object recognition and pose estimation even when the object is zoomed in, by partitioning the reference features sets and using metadata to optimize the image processing function.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a single reference features set is extracted from the original reference image, then the system is simple and computationally efficient, but object recognition and tracking are lost when the camera zooms in on the object
Solution Approach 1:
The reference image is divided into multiple slices, and a separate reference features set is extracted from each slice. This segmentation allows the system to maintain reliable object recognition and tracking even when zoomed in, as the appropriate slice-based reference features set can be selected based on the current view. The patent implements this by creating multiple reference features sets from different regions of the original reference image, enabling the system to handle various zoom levels effectively.
2Reliability
If multiple reference features sets are extracted from slices of the reference image, then object recognition is maintained at various zoom levels, but computational resources and processing complexity increase
Solution Approach 1:
The system dynamically selects which reference features set to use based on the current camera view and zoom level. Instead of constantly processing all reference features sets, the system adapts its processing to only the relevant slice-based reference features set, reducing computational energy consumption while maintaining reliable object recognition. This dynamic selection is implemented through a mapping function that determines the appropriate reference features set based on current image frame analysis.
Solution Approach 2:
Different regions (slices) of the reference image are processed to create specialized reference features sets optimized for specific areas. This allows the system to focus computational resources on locally relevant features rather than processing the entire reference image uniformly, reducing overall computational energy while maintaining recognition reliability in the current view.
3Measurement precision
If the camera zooms in on the object, then detailed view is achieved, but insufficient reference features are detected leading to loss of tracking
Solution Approach 1:
By segmenting the reference image into multiple slices and creating separate reference features sets for each slice, the system ensures that sufficient reference features are available even when the camera zooms in on a specific region. The slice-based approach guarantees that the relevant regional features are properly captured and can be matched against the current detailed view, maintaining tracking reliability at various zoom levels.
Data Source
AI summary
Method and systems for generating reference features sets for slices of a reference image. The reference features sets generated from slices enables better object recognition and/or tracking when a camera image only shows a portion of the reference image. Metadata is used to link the reference features set of the original image and of the slices together as belonging to the same object, providing hierarchical relationship information and/or spatial relationship information. An image processing function may be dynamically configured on the basis of whether an object has been successfully detected and the metadata associated with the object.