Printing Media Classification Using Multi-Model Spectral Discrimination
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing technologies for discriminating printing media using machine learning models face challenges in accurately distinguishing media with similar optical characteristics and lack effective management of training data and discrimination accuracy.
Innovation Solution
A method and system that utilize multiple machine learning models to classify spectral data of printing media, including a preparation step for machine learning models, acquisition of target spectral data, and execution of class classification processes to accurately discriminate printing media, with features like learning state determination and accuracy evaluation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a single machine learning model is used for discrimination, then the device complexity is low, but the discrimination precision for printing media with similar optical characteristics is insufficient
Solution Approach 1:
The patent divides the single discrimination task into multiple specialized machine learning models, where each model is trained to discriminate specific types of printing media. This segmentation allows each model to focus on particular characteristics, improving overall discrimination precision while managing complexity through modular architecture.
Solution Approach 2:
The patent combines multiple machine learning models into a unified discrimination system that integrates their outputs. By merging the results from multiple specialized models, the system achieves higher discrimination precision for printing media with similar optical characteristics while maintaining systematic management of the combined models.
2Measurement precision
If multiple machine learning models are prepared for accurate discrimination, then the discrimination precision improves, but the device complexity increases
Solution Approach 1:
The patent creates a universal discrimination system that can handle multiple types of printing media through multiple machine learning models. Each model serves multiple purposes: discrimination, learning state determination, and accuracy evaluation. This multi-functionality reduces the need for separate systems for each function, managing complexity while maintaining high precision.
Solution Approach 2:
The patent changes the parameter of model quantity from one to multiple, and introduces parameters for learning state and accuracy evaluation. By systematically managing these parameters through unified control mechanisms, the system achieves high discrimination precision while preventing complexity from becoming unmanageable.
3Measurement precision
If training data is continuously accumulated for machine learning, then the learning accuracy improves, but the management complexity of training data increases
Solution Approach 1:
The patent implements feedback mechanisms where the discrimination results and accuracy evaluations are fed back into the training data management system. This feedback loop automatically identifies which training data needs to be updated or added, improving learning accuracy while reducing management complexity through automated decision-making processes.
Solution Approach 2:
The system performs self-service in managing training data by automatically determining learning states and evaluating accuracy without external intervention. The machine learning models themselves identify when they need retraining and what data to use, reducing the complexity of external data management while continuously improving learning accuracy.
4Reliability
If the discrimination process is executed without managing learning states, then the process is simple, but the reliability of discrimination results deteriorates
Solution Approach 1:
The patent performs preliminary actions by determining the learning state of each machine learning model before executing discrimination. This preliminary check ensures that only models with sufficient learning are used, improving reliability while keeping the added process complexity minimal through straightforward state verification.
Solution Approach 2:
The system uses feedback from learning state determination and accuracy evaluation to improve the reliability of discrimination results. By continuously monitoring and adjusting based on performance feedback, the system maintains high reliability without requiring complex manual intervention processes.
Data Source
AI summary
A method for executing a discrimination process of a printing medium includes a step (a) of preparing N machine learning models when N is an integer of 1 or more, a step (b) of acquiring target spectral data which is a spectral reflectance of a target printing medium, and a step (c) of discriminating a type of the target printing medium by executing a class classification process of the target spectral data using the N machine learning models.