Image Processing Device Object Similarity Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional image processing techniques face challenges in accurately determining whether objects in images are the same, especially when the target area cannot be detected due to object posture or positional relations, and when images are not temporally close.
Innovation Solution
An image processing device that detects multiple predefined areas within objects in two images, extracts feature vectors from these areas, and determines object similarity using these vectors, allowing for accurate identification without temporal alignment and improving determination accuracy by considering multiple viewpoints.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If a single specific area (e.g., face) is used for object identification, then the identification process is simple, but determination cannot be made when the target area cannot be detected due to posture or positional relation
Solution Approach 1:
The patent divides the object into multiple predefined areas (e.g., head, body, limbs) and extracts feature vectors from each area independently. This segmentation allows the system to use multiple areas for identification instead of relying on a single area, thereby improving reliability when one area cannot be detected due to posture or positioning issues.
2Measurement precision
If multiple areas are detected and feature vectors are extracted from each, then determination accuracy improves, but the complexity of the processing system increases
Solution Approach 1:
The system segments the object into multiple predefined areas and extracts feature vectors from each area independently. This modular approach allows the complexity to be managed through standardized processing steps for each area, while the overall accuracy improves by aggregating information from multiple segments.
Solution Approach 2:
The patent combines feature vectors from multiple areas to make the final determination. By merging the results from different areas (head, body, limbs), the system achieves higher determination accuracy while maintaining a manageable processing architecture through systematic combination rules.
3Adaptability or versatility
If conventional tracking techniques are used, then temporal proximity is assumed, but cannot be applied when images are not temporally close
Solution Approach 1:
The patent creates a universal identification system that works regardless of temporal proximity between images. By extracting and comparing feature vectors from multiple predefined areas, the system can reliably identify objects even when images are taken at different times, making the technique applicable to both temporal and non-temporal image pairs.
Data Source
AI summary
An image processing device includes one or more processors. The processors detects two or more first partial areas corresponding to each of two or more portions among a plurality of portions that are included in an object and that are set in advance from a first image, and detect two or more second partial areas corresponding to each of two or more portions among the portions from a second image. The processors extracts two or more first feature vectors from two or more of the first partial areas, and extract two or more second feature vectors from two or more of the second partial areas. The processors determines whether an object included in the first image and an object included in the second image are same, by using the first feature vectors and the second feature vectors.