Image Signature Extraction Device Dynamic Feature Switching
Solution Overview
Problem
Existing image signature extraction methods face challenges in discriminating between images when the number of dimensions is increased, leading to larger image signature sizes and increased matching costs, while also complicating the matching method.
Innovation Solution
An image signature extraction device that includes first and second feature extraction means, a feature type determination means, and a feature vector generation means to analyze and determine the effectiveness of feature vectors, allowing for the combination of different feature types across dimensions to improve discrimination capability without increasing signature size.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the number of dimensions is increased to improve discrimination capability, then the types of images that can be discriminated are increased, but the size of the image signature increases
Solution Approach 1:
The image is divided into multiple blocks, and each block is processed independently to extract local features. This segmentation allows the system to capture fine-grained details without requiring a proportional increase in overall signature size, as each block contributes only its essential local characteristics to the final signature.
Solution Approach 2:
Different feature extraction methods are applied to different blocks based on their local characteristics. The system adapts the feature extraction strategy to each local region, using methods suited to the specific content of each block rather than applying a uniform approach across the entire image.
2Measurement precision
If the number of dimensions is increased to improve discrimination capability, then the types of images that can be discriminated are increased, but the matching cost increases
Solution Approach 1:
The patent extracts only the most essential and discriminative features from each image block, rather than processing all possible features. This selective extraction reduces the dimensionality of the signature while maintaining discrimination capability, thereby lowering the computational cost of matching operations.
Solution Approach 2:
The system dynamically adjusts feature extraction parameters and dimensionality based on the characteristics of the image blocks. By adapting the number of dimensions and feature types to the actual content needs, the system avoids unnecessary computational overhead while preserving discrimination performance.
3Measurement precision
If the number of dimensions is increased to improve discrimination capability, then the types of images that can be discriminated are increased, but the matching method becomes complicated
Solution Approach 1:
By dividing the image into independent blocks and processing each block separately, the system simplifies the matching process. Each block's features can be compared independently, and the overall match result is derived from combining these simpler local comparisons, avoiding the complexity of analyzing the entire high-dimensional signature at once.
Solution Approach 2:
The patent employs a uniform feature extraction and quantization framework across all image blocks. This homogeneous approach ensures that the same processing steps and comparison methods are applied consistently to each block, simplifying the overall matching procedure despite the presence of multiple feature types and dimensions.
Data Source
AI summary
The image signature extraction device includes a first feature extraction means for extracting, from an image, first features corresponding to the respective dimensions of a feature vector; a second feature extraction means for extracting, from the image, second features which are different from the first features, corresponding to the respective dimensions of a feature vector; a feature type determination means for analyzing at least one of the image and the extracted first features as a subject for analysis to determine whether or not the feature vector constituted of the extracted first features has effectiveness in discriminating an image, and if the feature vector has the effectiveness, determining the first features to be the type of the features used for the respective dimensions, while if the feature vector does not have the effectiveness, determining the second feature to be the type of the features used for at least part of the dimensions and determining the first features to be the type of the features used for the remaining dimensions; and a feature vector generation means for generating a feature vector of the image from the extracted first features and the extracted second features, according to the determined type of the features used for the respective dimensions.