Image Defect Prediction Using Spatial Frequency Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current systems for predicting defective images formed by image forming apparatuses struggle to accurately determine density irregularities, leading to excessive inspection time and material wastage, as they lack the ability to differentiate between conspicuous and inconspicuous irregularities and do not utilize original-image data effectively for prediction.
Innovation Solution
A system that analyzes image data by dividing it into regions of interest based on observation distance, calculates the spatial frequency of gradient distribution, and uses a correlation index to predict the conspicuousness of density irregularities, allowing for informed correction of image forming conditions and threshold setting.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If one hundred percent inspection is carried out to ensure quality, then accuracy of detecting defective images is improved, but loss of time and productivity deteriorate
Solution Approach 1:
The system performs preliminary analysis of original-image data before actual image formation to predict the occurrence of conspicuous density irregularities. By evaluating spatial frequency characteristics and gradient distributions in advance, the system identifies which images are likely to be defective, enabling selective inspection rather than exhaustive one hundred percent inspection.
Solution Approach 2:
The system introduces an intermediary prediction mechanism that bridges original-image data and final image quality assessment. By analyzing intermediate characteristics such as spatial frequency components and gradient distributions from the original image data, the system predicts conspicuousness of density irregularities before they manifest in the final printed image, allowing intelligent decision-making about which images require inspection.
2Device complexity
If conventional detection devices are used without prediction, then device complexity is reduced, but loss of information deteriorates due to inability to predict defective images
Solution Approach 1:
The system performs preliminary analysis of original-image data before actual image formation to predict the occurrence of conspicuous density irregularities. By evaluating spatial frequency characteristics and gradient distributions in advance, the system identifies which images are likely to be defective, enabling selective inspection rather than exhaustive one hundred percent inspection.
Solution Approach 2:
The system introduces an intermediary prediction mechanism that bridges original-image data and final image quality assessment. By analyzing intermediate characteristics such as spatial frequency components and gradient distributions from the original image data, the system predicts conspicuousness of density irregularities before they manifest in the final printed image, allowing intelligent decision-making about which images require inspection.
3Measurement precision
If threshold for density irregularity is set too low, then measurement precision is improved, but loss of substance deteriorates due to wasteful discarding of non-defective materials
Solution Approach 1:
The system applies different evaluation criteria and threshold settings to different regions of interest within images. By analyzing spatial frequency characteristics and gradient distributions locally in specific regions rather than applying a uniform threshold across the entire image, the system accurately identifies conspicuous density irregularities while preserving non-defective areas, thereby reducing wasteful discarding of good materials.
4Loss of substance
If threshold for density irregularity is set too high, then loss of substance is reduced, but measurement precision deteriorates due to undetected defective images
Solution Approach 1:
The system applies different evaluation criteria and threshold settings to different regions of interest within images. By analyzing spatial frequency characteristics and gradient distributions locally in specific regions rather than applying a uniform threshold across the entire image, the system accurately identifies conspicuous density irregularities while preserving non-defective areas, thereby reducing wasteful discarding of good materials.
Data Source
AI summary
A system for predicting occurrence of a defective image includes an observation distance obtainer which obtains an observation distance of an image to be formed by an image forming apparatus. The system (i) divides image data input to the image forming apparatus as an original of the image into regions of interest having a size determined based on the observation distance, (ii) analyzes a spatial frequency of a gradient distribution of the image with respect to each of the regions of interest, and (iii) calculates a probability of a target density irregularity being conspicuous in the image to be formed by the image forming apparatus based on the image data by using a correlation index between a result of the analysis and an evaluation value of the density irregularity.