Image Matching Using Quantized Frequency Features
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image matching techniques face challenges in achieving high-speed matching due to the complexity of inverse Fourier transform and the need for accurate representation of normalized cross power spectra, which requires significant computational resources.
Innovation Solution
An image matching apparatus that acquires frequency features of images, synthesizes them to generate a quantized synthesized frequency feature represented by binary or ternary values, and calculates a score indicating the degree to which this feature resembles a square wave having a single period, allowing for high-speed matching.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If inverse Fourier transform is applied to synthesized frequency features for image matching, then matching accuracy is improved, but computational complexity increases significantly
Solution Approach 1:
The patent extracts only the essential characteristic (single-period square wave pattern) from the complete frequency domain representation, eliminating the need for full inverse Fourier transform while preserving matching accuracy. This is achieved by directly analyzing the normalized cross power spectrum in the frequency domain and detecting its characteristic pattern.
Solution Approach 2:
The patent replaces the computationally expensive inverse Fourier transform with a simpler, lighter computational approach that directly evaluates the single-period square wave characteristic. This substitution maintains matching functionality while dramatically reducing computational burden.
2Measurement precision
If many bits are allocated to represent elements of normalized cross power spectrum for accurate matching, then matching precision is improved, but computational complexity increases
Solution Approach 1:
The patent changes the representation parameter from high-precision multi-bit floating-point numbers to simple binary values (+1/-1) indicating the presence or absence of single-period square wave characteristics. This parameter transformation maintains matching precision while drastically reducing computational complexity.
3Measurement precision
If traditional image matching methods are used, then matching accuracy is maintained, but processing speed decreases
Solution Approach 1:
The patent performs preliminary extraction of the single-period square wave characteristic from the normalized cross power spectrum before final matching determination. This preliminary action simplifies the subsequent matching process and enables faster processing while maintaining accuracy.
Data Source
AI summary
An image matching apparatus matching a first image against a second image includes an acquiring unit, a generating unit, and a determining unit. The acquiring unit acquires a frequency feature of the first image and a frequency feature of the second image. The generating unit synthesizes the frequency feature of the first image and the frequency feature of the second image, and generates a quantized synthesized frequency feature in which a value of an element is represented by a binary value or a ternary value. The determining unit calculates a score indicating a degree to which the quantized synthesized frequency feature is a square wave having a single period, and matches the first image against the second image based on the score.