Sheet Identification Apparatus Using Machine Learning Estimation Model
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing sheet type identification methods using optical sensors are prone to inaccuracies due to individual sensor differences and degradation over time, leading to insufficient identification accuracy when collating sensor output values with pre-specified values.
Innovation Solution
An identification apparatus employing machine learning to create an estimation model using sensor data from sheets, where the type of sheet is identified based on input parameters, and re-learning is performed to adjust the model when discrepancies arise, ensuring accurate identification by updating the model with user-inputted sheet types.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If collation with pre-specified values is used for sheet type identification, then the identification process is simple, but identification accuracy deteriorates due to sensor individual differences and degradation
Solution Approach 1:
The system performs preliminary learning to create an estimation model before actual sheet type identification. Sensor data from multiple sheets of known types are collected and used to train the machine learning model, establishing a baseline that accounts for individual sensor characteristics before deployment
Solution Approach 2:
The system transitions from using fixed pre-specified threshold values to using dynamic estimation models that adapt parameters based on actual sensor data. The identification criteria change from static collation against predetermined values to dynamic probability-based classification using the trained model
2Measurement precision
If machine learning estimation model is used for sheet type identification, then identification accuracy is improved, but device complexity increases
Solution Approach 1:
The system performs self-calibration through automated learning processes. The machine learning model automatically adjusts its parameters by learning from sensor data, eliminating the need for manual calibration and reducing operational complexity despite the increased initial system complexity
Solution Approach 2:
The system incorporates feedback mechanisms where identification results and user corrections are fed back into the learning model. When users correct misidentifications, this feedback is used to re-train and improve the model, creating a self-improving system that reduces long-term operational complexity
3Measurement precision
If re-learning is performed to correct identification errors, then identification accuracy is maintained, but processing time increases
Solution Approach 1:
Instead of performing complete re-learning from scratch, the system performs incremental or partial re-learning using only the new feedback data. This selective updating approach maintains accuracy while minimizing the time and computational resources required compared to full re-training
Solution Approach 2:
The system performs preliminary learning during manufacturing or initial setup to establish the base model. This preliminary action reduces the need for extensive re-learning later, as the model already has a strong foundation that requires only minor adjustments through feedback
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
This approach enhances sheet type identification accuracy by dynamically updating the estimation model based on user feedback, accounting for sensor variations and improving identification performance across different sheet types.
Implementation Method 1
the sheet is illuminated with light, and sheet type identification is performed based on output values obtained when a sensor receives reflected light and transmitted light
Data Source
AI summary
An identification apparatus uses a sensor to sense a sheet, thereby obtaining a parameter concerning a characteristic of the sheet, and identifies a type of the sheet based on a result of inputting the parameter obtained by the sensor to an estimation model obtained by machine learning using a parameter of a sheet as input data and a type of a sheet as supervised data, the type used as supervised data corresponding to the parameter used as input data. The apparatus performs re-learning such that the identified type of the sheet based on a result of inputting the parameter obtained by the sensor as the input data to the estimation model becomes the type of the sheet represented by input information if the identified type of the sheet and the type of the sheet represented by the input information are different.