Image Matching Using Texture Filtering for Object Recognition
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current image processing techniques for object recognition in images face challenges with speed and accuracy, particularly in real-time and batch processing of multiple images, due to noise, texture, and repeating patterns, which can lead to false positives and negatives.
Innovation Solution
An image processing system that uses feature point extraction and matching algorithms, such as SIFT and SURF, combined with filters and segmentation methods to reduce the impact of noise and texture, and employs a database system to efficiently match template images with test images, allowing for real-time and batch processing.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If traditional image processing techniques are used for object recognition, then the system can process images, but the speed and accuracy are insufficient particularly in real-time and batch processing
Solution Approach 1:
The patent segments the image processing task into multiple independent stages: feature point extraction, descriptor generation, filtering (removing low-scoring feature points), and matching. This segmentation allows parallel processing of different image regions and feature types, improving processing speed while maintaining accuracy through specialized handling at each stage
Solution Approach 2:
The patent performs preliminary actions by pre-computing and storing descriptors for template images in a database before actual matching occurs. During real-time processing, only feature extraction and comparison are needed, significantly reducing processing time while maintaining recognition accuracy through the pre-prepared reference data
2Reliability
If traditional matching algorithms are used, then object recognition can be performed, but false positives and negatives increase due to noise, texture, and repeating patterns
Solution Approach 1:
The patent extracts only the most distinctive and reliable feature points from images, filtering out feature points with low descriptor scores that are likely to be caused by noise, texture, or repeating patterns. This selective extraction removes harmful factors while preserving the essential matching information
Solution Approach 2:
The patent applies different processing quality levels to different feature points based on their local characteristics. High-scoring feature points with distinctive local patterns receive full processing attention, while low-scoring points from noisy or repetitive regions are filtered out, optimizing both accuracy and processing efficiency
3Measurement precision
If comprehensive feature matching is performed on all images, then recognition accuracy can be maintained, but processing time increases significantly for batch processing
Solution Approach 1:
The patent performs partial matching by processing only the most promising feature point pairs identified through initial filtering and scoring. Instead of exhaustively comparing all feature points, the system focuses computational resources on high-probability matches, achieving sufficient accuracy with significantly reduced processing time for batch operations
Data Source
AI summary
A computer-implemented method for determining whether a first image contains at least a portion of a second image includes: determining a first set of feature points associated with the first image; removing from said first set of feature points at least some feature points in the first set that correspond to one or more textures in the first image; and then attempting to match feature points in said first set of feature points with feature points in a second set of feature points associated with said second image to determine whether said first image contains at least a portion of said second image.