Object Identification Using Geometric Transformation and Local Descriptors
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional object identification systems using local descriptors often incorrectly identify images showing different objects due to the presence of numerous corresponding feature points, especially when the differences between objects are minor.
Innovation Solution
An object identification apparatus and method that includes a local descriptor matching unit to determine correct correspondences between feature points in input and reference images, an input image difference area descriptor extracting unit to correct geometric deviations, and a descriptor matching unit to output matching results, allowing for more accurate identification by focusing on specific areas with geometric transformations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If local descriptors are extracted from brightness information and compared to identify objects, then the identification process can be performed, but identification accuracy deteriorates when objects differ only slightly
Solution Approach 1:
The patent segments the image into multiple local areas around feature points and extracts descriptors from each segment. By comparing descriptors from corresponding local areas rather than using global descriptors, the system can detect subtle differences between objects while maintaining identification capability. This segmentation approach allows precise local comparison that reveals differences even when overall object appearance is similar.
Solution Approach 2:
The patent applies local quality by extracting and comparing descriptors from specific local areas around feature points rather than using global image properties. Each local area is processed independently to generate descriptors that capture local characteristics, enabling the system to distinguish objects based on local differences while ignoring global similarities. This local-focused approach improves accuracy for objects with minor differences.
2Measurement precision
If geometric transformation is applied to correct geometric deviation, then identification accuracy improves, but processing complexity increases
Solution Approach 1:
The patent performs geometric transformation as a preliminary action before descriptor extraction and comparison. By correcting geometric deviation in advance through transformation of the input image or reference image, the system prepares the data in a standardized format that facilitates more accurate subsequent descriptor matching. This preliminary geometric correction simplifies the overall identification process while improving accuracy.
Solution Approach 2:
The patent changes geometric parameters (position, orientation, scale) through transformation to align images before descriptor extraction. By adjusting these parameters to correct geometric deviation, the system creates a standardized representation that improves descriptor comparison accuracy. This parameter transformation approach handles geometric variations without adding complex processing steps during the core identification algorithm.
Data Source
AI summary
An input image showing a same object as an object shown in a reference image is identified more accurately. A difference area in the input image is determined by converting a difference area in the reference image, on a basis of geometric transformation information calculated by an analysis using a local descriptor. By matching a descriptor extracted from the difference area in the input image with the difference area in the reference image, fine differences that cannot be identified by conventional matching using only a local descriptor can be distinguished and images showing a same object can be exclusively identified.