Machine-Learned Color Profile Creation for Printing Apparatus
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for creating color conversion profiles in printing apparatuses require extensive man-hours and repeated printing and imaging processes to set upper limit values for ink discharge, leading to inefficiencies and high labor costs.
Innovation Solution
An information processing apparatus that utilizes a machine-learned model to estimate the limit value of ink usage per unit area, reducing the need for repeated printing and imaging by integrating a learning apparatus for supervised machine learning, which generates a color conversion profile with improved color reproducibility.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If repeated printing and imaging processes are performed to determine upper limit values for each effect, then printing quality is improved, but man-hours and time consumption increase significantly
Solution Approach 1:
The system performs preliminary actions by collecting printing condition information and effect information in advance, then uses machine learning to pre-determine upper limit values for multiple effects simultaneously. This eliminates the need for repeated printing and imaging processes for each individual effect, significantly reducing time consumption while maintaining printing quality.
Solution Approach 2:
The system creates a virtual model through machine learning that copies and simulates the complex relationships between printing conditions, effects, and upper limit values. Instead of physically repeating printing and imaging processes, the machine learning model predicts the outcomes based on learned patterns from training data,大幅 reducing time and labor requirements.
2Manufacturing precision
If multiple separate determinations are made for different effects (overflowing, bleeding, aggregation, color saturation), then comprehensive quality control is achieved, but process complexity increases
Solution Approach 1:
The system merges multiple separate determination processes into a single integrated machine learning model. The model simultaneously processes printing condition information and effect information to determine upper limit values for all effects (overflowing, bleeding, aggregation, color saturation) in one operation, reducing process complexity while maintaining comprehensive quality control.
Solution Approach 2:
The machine learning model serves as a universal tool that handles multiple determination functions simultaneously. Instead of having separate processes for each effect, the single model can predict upper limit values for all effects based on input printing conditions, simplifying the overall system while maintaining comprehensive control.
3Measurement precision
If threshold values from sensory evaluation are used for determination, then subjectivity is reduced, but the requirement for multiple test patterns and repeated processes remains
Solution Approach 1:
The system replaces the mechanical process of sensory evaluation with a machine learning-based computational system. The machine learning model objectively determines upper limit values by learning from training data, eliminating the need for multiple test patterns and repeated sensory evaluations while maintaining or improving determination accuracy.
Solution Approach 2:
The system changes the determination approach from using fixed threshold values based on sensory evaluation to using dynamic predictions from a machine learning model. The model adjusts upper limit values based on learned relationships between printing conditions and effects, improving both accuracy and efficiency simultaneously.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
An information processing apparatus includes: a storage section storing a machine-learned model that learned, by machine learning, a relationship between a type of a printing medium, an amount of a coloring material on the printing medium per unit area, and an image printed on the printing medium; a receiving section receiving an input of selection information including medium-type information regarding a type of the printing medium; an obtaining section obtaining imaging information obtained by capturing the image printed on the printing medium by a printing section performing printing by using the coloring material; an estimating section estimating, based on the selection information and the imaging information, by using the machine-learned model a limit value indicating a maximum value or a minimum value of an amount of the coloring material to be used in printing on the printing medium by the printing section per unit area; and a creating section creating, by using the limit value, a color conversion profile including information regarding mapping between a coordinate value in a color space and an amount of the coloring material.