Image Recognition via Feature Restoration and Matching
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image recognition methods may incorrectly identify images as specific images due to the presence of numerous similar image features, leading to inaccurate recognition even if the input image is visually distinct from the specific image.
Innovation Solution
An image recognition method and device that detect and store relevant image features from a specific image, generate a restoration image using selected features from an input image, and determine matching through a matching process to accurately recognize the specific image by using the portions and feature intensities detected from the specific image.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If image recognition is performed based on whether each image feature of a specific image is included, then the recognition process is simple, but recognition accuracy deteriorates when a large number of the same image features are included in different images
Solution Approach 1:
The patent divides the image into multiple blocks and extracts features from each block separately. Instead of treating the entire image as a single unit, the image is segmented into manageable portions (e.g., 8x8 pixel blocks), and features are extracted from each segment. This segmentation allows the system to capture local patterns more effectively while reducing the impact of globally similar features across different images.
Solution Approach 2:
The patent transforms the recognition approach by adding a spatial dimension to feature comparison. Instead of simply counting the presence of features, the system considers the spatial arrangement and relationships between features across different image blocks. This dimensional transformation enables differentiation between images that have the same feature set but different spatial configurations.
2Productivity
If only image features are compared without spatial arrangement consideration, then the processing is efficient, but images with similar feature sets but different arrangements are erroneously recognized as identical
Solution Approach 1:
The patent segments the image into multiple blocks and processes each block independently to extract features. This segmentation maintains processing efficiency by allowing parallel feature extraction from different blocks while capturing local spatial patterns within each block, thereby improving reliability without significantly increasing overall processing complexity.
Solution Approach 2:
The patent applies local quality analysis by examining the specific characteristics and arrangements of features within each image block separately. Each block's feature arrangement is analyzed in its local context, allowing the system to capture nuanced spatial relationships that distinguish similar images while maintaining efficient processing through localized analysis.
Data Source
AI summary
A method of recognizing a specific image includes detecting image features from the specific image and storing in a memory each of detected relevant image features, detecting relevant image features from an input image and selecting from the memory a portion corresponding to each of the detected relevant image features, generating a restoration image using portions selected in the selecting, and determining whether the generated restoration image matches with the input image by a matching process and recognizing that, when it is determined that the restoration image matches with the input image, the input image is the specific image.