Object Detection Metadata Coordinate Transformation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current technologies face challenges in accurately detecting and tracking objects within images and video frames, particularly in maintaining object location and identity across different frames and coordinate systems, especially when aspect ratios change or objects overlap.
Innovation Solution
Generating object detection metadata that includes perimeter coordinates in a first coordinate system, converting them to a second system with the same aspect ratio, and adding layering values to indicate depth and unique index values for object matching, along with weighting probabilities for face detection and matching.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If object detection is performed in pixel coordinates, then detection accuracy is maintained, but compatibility across different image resolutions and aspect ratios deteriorates
Solution Approach 1:
The patent introduces a new dimension of coordinate representation by converting from pixel-based coordinates to normalized coordinates. Normalized coordinates use the range [0, 1] where 0 represents the top-left corner and 1 represents the bottom-right corner of the image. This dimensional transformation allows the same detection results to be applied across images of different resolutions and aspect ratios, resolving the contradiction between maintaining detection accuracy and achieving cross-resolution compatibility.
2Quantity of substance
If multiple objects are detected in overlapping regions, then detection completeness is improved, but identification accuracy deteriorates
Solution Approach 1:
The patent segments the detection results by assigning unique object IDs to distinguish between different objects even when they overlap in the image. Each detected object receives a unique identifier that tracks it across multiple frames, allowing the system to maintain both the completeness of detecting multiple objects and the accuracy of identifying each individual object correctly.
Solution Approach 2:
The patent implements feedback mechanisms through confidence scores and matching algorithms that continuously refine object identification. The system uses feedback from detection results to adjust and improve identification accuracy, particularly in overlapping regions where multiple objects are present, ensuring that correct objects are identified even when spatial overlap occurs.
3Measurement precision
If face detection is performed across multiple frames, then tracking accuracy is improved, but computational complexity increases
Solution Approach 1:
The patent applies preliminary action by pre-processing frames to extract key features and characteristics before performing the full detection and tracking algorithm. This includes pre-computing normalized coordinates, extracting dominant colors, and identifying key facial features in advance, which reduces the computational burden during the actual tracking process across multiple frames while maintaining high accuracy.
Data Source
AI summary
A perimeter around a detected object in a frame of image data can be generated in a first coordinate system. The perimeter can be converted from the first coordinate system into a second coordinate system having the same aspect ratio as the first coordinate system. A first metadata entry can include dimensions of image data in the second coordinate system. A second metadata entry can provide a location and dimensions of the converted perimeter in the second coordinate space. Additional metadata can indicate matching objects between frames, position of an object relative to other objects in a frame, a probability that an object is correctly detected, and a total number of objects detected across multiple frames of image data.