Print Medium Specification via Segmented ML Discrimination
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing print medium specification methods using machine learning models require extensive time for learning when multiple physical property information are used, leading to inefficiencies in discriminating the type of a print medium.
Innovation Solution
A method and system that acquire first and second physical property information about a print medium, input the first information into a learned machine learning model to generate discrimination information, and use this information along with the second information not used for learning to specify the print medium type, employing a vector neural network for efficient classification.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If multiple physical property information are used for machine learning, then discrimination accuracy is improved, but learning time increases
Solution Approach 1:
The patent segments the physical property information into two distinct groups: first physical property information used for machine learning model construction, and second physical property information reserved for runtime specification. This segmentation allows the system to train on comprehensive data while making efficient use of resources during actual operation, resolving the contradiction between using multiple parameters for accuracy and the time cost of processing all of them.
Solution Approach 2:
The patent extracts and separates the second physical property information from the machine learning process entirely. By taking out this information and using it only for final specification determination rather than training, the system achieves high discrimination accuracy without incurring the full time cost of processing all available physical property data through the learning model.
2Measurement precision
If a learned machine learning model is used for discrimination, then discrimination accuracy is improved, but processing time increases
Solution Approach 1:
The patent applies partial action by using the machine learning model only for processing the first physical property information, while handling the second physical property information through direct comparison with stored specification data. This partial use of the learned model maintains high discrimination accuracy while avoiding the excessive processing time that would result from applying the model to all available information.
Solution Approach 2:
The patent performs preliminary machine learning training during the model construction phase, storing the learned discrimination function for later use. During runtime specification, the pre-trained model quickly processes the first physical property information without requiring re-learning, thereby maintaining high accuracy while improving processing efficiency during actual operation.
Data Source
AI summary
An print medium specification method includes (a) step for acquiring first physical property information related to the print medium; (b) step for acquiring second physical property information different from the first physical property information related to the print medium; (c) step for acquiring a discrimination information for discriminating the type of the print medium by inputting the first physical property information to a discrimination function configured as a learned machine learning model; and (d) step for specify a type of the print medium using the discrimination information and the second physical property information not used for machine learning.