Image Feature Extraction Using Gradient Covariance Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional image recognition devices face challenges in accurately extracting feature data that effectively distinguishes recognition target objects from others, particularly in capturing the statistical information necessary for high classification accuracy.
Innovation Solution
An image feature extraction device that calculates intensity gradient data in two directions, divides the image into partial areas, computes covariance matrices, and outputs feature vectors based on the parameters of ellipses derived from these matrices, using quantized logarithmic values to enhance recognition accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional feature extraction methods are used, then the extraction process is simple, but the recognition accuracy is insufficient due to lack of statistical information
Solution Approach 1:
The image is divided into multiple local areas, and gradient distributions are calculated separately for each area. This segmentation allows the extraction of localized statistical features that improve recognition accuracy while maintaining a systematic approach to feature extraction.
Solution Approach 2:
The invention transforms gradient data into statistical parameters by calculating covariance matrices and extracting eigenvalues. This parameter transformation converts raw gradient information into meaningful statistical features that enhance recognition accuracy.
2Measurement precision
If detailed gradient analysis is performed for high accuracy, then recognition performance improves, but computational complexity increases
Solution Approach 1:
The invention extracts only the essential statistical features (eigenvalues of covariance matrices) from the gradient data, discarding redundant information. This extraction approach achieves high recognition accuracy while reducing computational power requirements by focusing only on the most discriminative features.
Solution Approach 2:
Instead of analyzing all gradient details uniformly across the entire image, the method applies gradient analysis selectively to local areas and extracts only the most significant statistical parameters. This partial action approach balances accuracy with computational efficiency.
Data Source
AI summary
An image feature extraction device according to an embodiment includes a gradient image calculator generates intensity gradient data with respect to two different directions based on intensity data of image data; and a gradient count unit calculates a covariance matrix for each partial area obtained by dividing the image data based on the intensity gradient data. The image feature extraction device according to the embodiment further includes a feature data output unit calculates two parameters related to a major axis and a minor axis of an ellipse expressed by the covariance matrix, quantizes a range of the logarithms of the parameters for each of the partial area using a predetermined division number, and outputs a feature vector which contains a value only at a dimension corresponding to the quantized range different from the other dimensions.