Subject Recognition via Feature Point Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image recognition methods struggle with accurately identifying occluded subjects in target images and require a large number of key frames, making them inefficient for mobile device implementation due to increased data volume.
Innovation Solution
A processor-implemented method that extracts feature points from target images, measures movement information, groups similar feature points, and determines subject types using an image vocabulary tree, allowing for recognition based on non-occluded portions of the subject without considering occluded areas.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a great number of key frames are considered for accurate subject recognition, then recognition accuracy is improved, but data volume increases making mobile device implementation difficult
Solution Approach 1:
The patent segments the image into multiple regions and extracts feature points from each region. By processing images in segments rather than as complete key frames, the system reduces data volume while maintaining recognition accuracy through the aggregated feature information from all segments.
Solution Approach 2:
The patent extracts only the essential feature points from images rather than using complete images as key frames. This extraction approach retains the critical information needed for subject recognition while significantly reducing the data volume required for processing.
2Reliability
If verification approaches using registered key frames are used, then subject recognition is enabled, but occluded subjects cannot be recognized or are falsely recognized
Solution Approach 1:
The patent applies local quality by analyzing different regions of the image independently through feature point extraction. Each region's feature points are processed separately, allowing the system to recognize subjects based on visible local characteristics even when parts of subjects are occluded, rather than requiring complete view of the entire subject.
Solution Approach 2:
The patent uses partial action by recognizing subjects based on partial visibility of their features. The system can identify subjects even when only portions are visible by extracting and analyzing feature points from the visible regions, eliminating the need for complete subject visibility that traditional key frame verification requires.
3Adaptability or versatility
If multiple subjects are included in a target image, then recognition complexity increases, but accuracy to recognize each subject individually decreases
Solution Approach 1:
The patent segments the image into multiple regions and extracts feature points from each region independently. This segmentation allows the system to process multiple subjects simultaneously by analyzing their respective regions separately, maintaining high accuracy for each subject even when multiple subjects are present in the same image.
Data Source
AI summary
Disclosed is a subject recognizing apparatus and method. The method may include extracting feature points from a target image, respectively measuring movement information of each of a plurality of the extracted feature points, selectively grouping the extracted feature points into one or more groups based on the respectively measured movement information, determining a type of subject present in at least one group of the one or more groups based on at least a portion of the subject present in the at least one group, and recognizing a subject included in the target image based on the determined type of subject.