Image Sharpness Detection Using Frequency Domain Normalization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image sharpness evaluation methods in the frequency domain face challenges when comparing images with different camera settings, lighting, and sensor resolutions, leading to inconsistent frequency information and incorrect sharpness characterization.
Innovation Solution
The method involves extracting regions of interest from images, normalizing them to a common scale, and using Discrete Cosine Transform (DCT) to evaluate frequency information, with multiple scales and patch sizes to assess local sharpness, and a multi-segment linear function to map DCT scores to sharpness scores, ensuring consistent evaluation across varying image sizes and settings.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If frequency domain methods are used to evaluate image sharpness, then sharpness detection capability is improved, but measurement consistency deteriorates when images have different camera settings, lighting, or sensor resolutions
Solution Approach 1:
The patent transforms images to a common frequency domain representation (e.g., DCT coefficients) and normalizes frequency scales across images with different resolutions and camera settings. This parameter transformation allows consistent sharpness evaluation by comparing frequency characteristics rather than spatial pixel values, resolving the contradiction between detection capability and measurement consistency.
2Measurement precision
If images are normalized to a common scale, then frequency information comparability is improved, but processing complexity increases due to multiple normalization steps
Solution Approach 1:
The patent performs preprocessing steps including resizing images to a standard dimension, converting to grayscale, and applying DCT transformation before sharpness evaluation. These preliminary actions standardize the input data format and frequency characteristics, enabling consistent comparison while managing complexity through systematic preprocessing rather than complex algorithms during evaluation.
3Measurement precision
If multiple scales and patch sizes are used to assess local sharpness, then evaluation accuracy is improved, but computational load increases
Solution Approach 1:
The patent divides images into multiple patches or regions of interest and evaluates sharpness at different scales (e.g., fine, medium, coarse frequencies) separately. This segmentation allows focused computation on local features rather than processing entire images uniformly, improving evaluation accuracy for different regions while managing computational load through selective multi-scale analysis.
Solution Approach 2:
The patent applies multi-scale analysis selectively to regions where it provides the most benefit, such as regions containing important features or areas with varying sharpness characteristics. Rather than uniformly applying complex multi-scale processing to entire images, the method uses partial action to achieve high accuracy where needed while reducing computational burden in less critical areas.
Data Source
AI summary
An image processing apparatus and method are provided which obtains an image captured by an image capture device and stored in a memory, extracts one or more regions of interest in the obtained image, normalizes the extracted one or more regions of interest to be at a same scale or some predefined scales, extract the frequency information of regions of interest, determines a sharpness of the obtained image by aggregating the frequency information in each of the one or more extracted regions of interest, and labels the obtained image with the determined sharpness score.