Printer Medium Type Detection Using Machine Learning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for determining the type of a print medium based on physical property values, such as reflectance and transmittance, are inefficient unless a suitable threshold value is set, making it difficult to accurately determine the medium's type.
Innovation Solution
A printer and machine learning device that utilize a machine-learned model, trained on data including reflectance, transmittance, and image data of the print medium, to determine its type, eliminating the need for manual threshold setting and improving determination accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If a threshold value-based method is used to determine print medium type, then the determination process is simple, but the determination rate deteriorates unless the threshold value is suitable
Solution Approach 1:
The patent transforms the determination approach from using fixed threshold values to using machine learning models that process multiple physical property parameters (reflectance, transmittance, surface roughness, grain) simultaneously. This allows the system to adaptively determine print medium types based on learned patterns from teaching data, improving determination accuracy without requiring manual threshold optimization for each medium type.
Solution Approach 2:
The patent uses image data obtained by reading the surface of the print medium as a copy or representation of the medium's physical characteristics. This image data is processed through machine learning models to infer the print medium type, replacing the need for direct physical measurement and threshold comparison while maintaining determination capability.
2Adaptability or versatility
If manual threshold determination is used for each print medium type, then the determination condition can be customized, but it is not easy to determine suitable threshold values
Solution Approach 1:
The patent implements a self-service mechanism where the machine learning model automatically learns optimal determination conditions from teaching data that includes various print medium types and their corresponding physical property values. The system performs self-learning and self-optimization without requiring users to manually determine or adjust threshold values for each medium type, thereby maintaining adaptability while reducing complexity.
Solution Approach 2:
The patent performs preliminary action by pre-training the machine learning model with teaching data that encompasses multiple print medium types and their characteristics. This pre-learning process prepares the model with optimal determination conditions before actual use, eliminating the need for users to perform complex threshold determination work during operation.
3Measurement precision
If multiple physical property values are used in the machine learning model, then the possibility of correctly determining print medium type increases, but the data processing complexity increases
Solution Approach 1:
The patent replaces complex mechanical or manual data processing methods with machine learning algorithms that automatically process multiple physical property values (reflectance, transmittance, surface roughness, grain). The machine learning model handles the complexity of multi-parameter analysis through automated computation and pattern recognition, improving determination accuracy while managing processing complexity through algorithmic efficiency.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
The solution enables accurate and efficient determination of the print medium's type, reducing the risk of print quality deterioration by using a machine-learned model to analyze spectral reflectance and transmittance characteristics, thereby improving the reliability of print operations.
Implementation Method 1
reflectance of the print medium, transmittance of the print medium, and image data obtained by capturing an image of a surface of the print medium
Implementation Method 2
reflectance of the print medium, transmittance of the print medium, and image data obtained by capturing an image of a surface of the print medium
Data Source
AI summary
A printer includes: a memory configured to store a machine-learned model obtained by machine learning using teaching data associating at least one of reflectance of a print medium, transmittance of the print medium, and image data obtained by capturing an image of a surface of the print medium with a type of the print medium; and a print controller configured to determine a type of a print medium using at least one of reflectance of the print medium, transmittance of the print medium, and image data obtained by capturing an image of a surface of the print medium and the machine-learned model.