Image processing program, image processing method, detection program, detection method, and information processing device.
By standardizing image appearance through a transformation rule that aligns brightness values, the method addresses imaging condition variations, enhancing the accuracy of machine learning in image analysis.
JP7883166B2Active Publication Date: 2026-07-01FUJITSU LTD
Patent Information
- Authority / Receiving Office
- JP · JP
- Patent Type
- Patents
- Current Assignee / Owner
- FUJITSU LTD
- Filing Date
- 2022-11-04
- Publication Date
- 2026-07-01
AI Technical Summary
Technical Problem
Existing image analysis technologies face challenges in achieving accurate machine learning due to differences in imaging conditions, leading to variations in how images are presented, which hinder the effectiveness of machine learning models.
Method used
An image processing method that generates a transformation rule to align the brightness values of images captured under different conditions, ensuring similarity in histogram distributions, thereby improving the accuracy of machine learning by reducing variations in image appearance.
Benefits of technology
The method enhances the accuracy of machine learning by standardizing image appearance, allowing for effective use of images captured under varying conditions, thereby improving the precision of image analysis.
✦ Generated by Eureka AI based on patent content.
Smart Images
Figure 0007883166000001 
Figure 0007883166000002 
Figure 0007883166000003
Abstract
The present invention reduces the difference between methods for capturing images to be used for machine-learning. The information processing device generates a histogram that indicates the number of pixels for each brightness value of a first image of a first subject captured in a first image-capturing condition. In addition, the information processing device generates a histogram that indicates the number of pixels for each brightness value of a second image of a second subject of the same type as the first subject captured in a second image-capturing condition. Then, the information processing device generates a conversion rule, by which the similarity of the histogram of the first image and the histogram of the second image is improved, for the pixel brightness values of the second image. Then, the information processing device converts, according to the conversion rule, the brightness value of each pixel of a third image of a third subject of the same type as the first subject captured in the second image-capturing condition. In addition, the information processing device executes machine-learning using the third image of which the brightness values are converted.
Need to check novelty before this filing date? Find Prior Art