Image processing device, image processing method, and program

The image processing apparatus uses a machine learning model to enhance area discrimination between character and halftone regions, correcting misclassifications and improving image quality in complex characters, thus preventing quality deterioration.

JP2026087203APending Publication Date: 2026-05-27KONICA MINOLTA INC
View PDF 1 Cites 0 Cited by

Patent Information

Authority / Receiving Office
JP · JP
Patent Type
Applications
Current Assignee / Owner
KONICA MINOLTA INC
Filing Date
2024-11-15
Publication Date
2026-05-27

AI Technical Summary

Technical Problem

Existing image processing technologies struggle to accurately distinguish between character and halftone areas, leading to misclassification and uneven image quality, particularly in complex characters, which results in variations in character density and quality when converting to compressed formats.

Method used

An image processing apparatus utilizing a machine learning model for region discrimination to accurately differentiate between character and halftone regions, with additional features for correcting misclassifications and incorporating feedback mechanisms for improved learning.

Benefits of technology

Prevents deterioration in character quality by enhancing the accuracy of area discrimination, ensuring consistent image quality even in complex characters and reducing the need for rescanning.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure 2026087203000001_ABST
    Figure 2026087203000001_ABST
Patent Text Reader

Abstract

This invention provides an image processing device, an image processing method, and a program that can prevent a decrease in character quality. [Solution] An image processing device 1 performs image processing on an image in which region attributes are defined for each pixel by region discrimination that distinguishes between character regions and halftone regions, and comprises a machine learning discrimination unit 100 that performs region discrimination using a machine learning model 19. When performing region discrimination, the machine learning discrimination unit 100 compares the characters extracted from the input image with the machine learning model 19 and determines the region attribute of the characters as a character attribute.
Need to check novelty before this filing date? Find Prior Art

Citation Information

Patent Citations

  • Image processing apparatus, image processing method, and computer program

    JP2013123157A