Person Identification Using Multi-Resolution Feature Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image processing techniques face challenges in accurately determining whether a person in a first area and a person in a second area, where the second area overlaps with the first area but is of different size, are the same individual, especially in low-resolution images.
Innovation Solution
An image processing apparatus and method that detect a first area and a second area within an image, where the second area overlaps with the first area and is of different size, and estimate if the persons in both areas are the same by analyzing feature points and circumscribed shapes using convolutional neural networks (CNNs) for high accuracy without requiring high-resolution images.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If multiple area detection methods are used to improve identification accuracy, then measurement precision is improved, but device complexity increases
Solution Approach 1:
The patent divides the detection process into multiple specialized detection means, each targeting specific areas (face area, head area, body area) with different detection algorithms. This segmentation allows each detector to be optimized for its specific task, improving overall identification accuracy while maintaining manageable complexity through functional division.
Solution Approach 2:
The patent employs multiple detection methods that may detect overlapping or redundant areas. By using excessive detection actions (multiple detectors for the same or overlapping regions), the system ensures that identification accuracy is maintained even if some detectors fail or produce false results, resolving the contradiction between precision and complexity.
2Measurement precision
If high-resolution images are used to improve detection accuracy, then measurement precision is improved, but loss of energy increases
Solution Approach 1:
The patent applies different detection strategies to different local regions of the image. High-resolution detailed analysis is applied only to key areas (face, head) where identification is most critical, while other areas use coarser detection methods. This local differentiation maintains detection accuracy for critical features while reducing overall energy consumption compared to processing the entire image at high resolution.
Solution Approach 2:
The image processing is segmented into different resolution levels for different regions. By dividing the image into multiple areas and applying appropriate detection complexity to each segment, the system achieves high detection accuracy where needed while minimizing energy expenditure on less critical regions.
Data Source
AI summary
An image processing apparatus 1 comprises: a first area detecting means 11 for detecting a first area including at least a part of a person from an image, a first feature point detecting means 12 for detecting first feature points from the first area, a second area detecting means 13 for detecting a second area including at least a part of a person from the image, wherein the second area overlapping at least a part of the first area, wherein a size of the second area being different from the first area, a second feature point detecting means 14 for detecting second feature points from the second area, and an estimating means 15 for estimating whether or not a person included in the first area and a person included in the second area are the same person on the basis of the first feature points and the second feature points.