Image Forming Apparatus With Machine-Learned Show-Through Correction

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

The phenomenon of show-through occurs in image data obtained by reading devices, where the image on the backside of a document is captured during the scanning process, leading to unwanted interference.

Innovation Solution

An image forming apparatus utilizes a scanner to read documents, generates training data using machine learning, and performs show-through correction on read data using a trained model to mitigate this interference.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If a scanner reads a document, then image data is obtained, but show-through interference occurs where images on the back side are captured

Engineering Contradiction:

Improvereading accuracyVSAvoidshow-through interference

Core Design Contradiction:

Measurement precisionVSObject-affected harmful factors

Solution Approach 1:

The patent applies machine learning to convert the harmful show-through effect into a solvable pattern recognition problem. By training the model with paired images (front and back sides), the system learns to identify and eliminate show-through interference, transforming the harmful optical phenomenon into a correctable data pattern that improves reading accuracy

Inventive Principle:

Principle #22Blessing in disguise (Convert harm into benefit)

Solution Approach 2:

The patent changes the parameter space by introducing training conditions (light source characteristics, paper properties, document thickness) as inputs to the machine learning model. This allows the system to adapt correction parameters based on specific document characteristics, dynamically adjusting the correction process to maintain high reading accuracy across different document types

Inventive Principle:

Principle #35Parameter changes

2Reliability

If machine learning training processing is performed to generate a trained model, then show-through correction capability is improved, but processing time and computational resources increase

Engineering Contradiction:

Improveshow-through correction accuracyVSAvoidtraining processing time

Core Design Contradiction:

ReliabilityVSLoss of time

Solution Approach 1:

The patent performs machine learning training in advance to generate a pre-trained model before actual document scanning. This preliminary action separates the time-consuming training phase from the operational scanning phase, allowing the model to be trained once on comprehensive datasets while enabling rapid show-through correction during subsequent document reading operations

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The patent implements a two-stage approach where a general show-through correction model is trained on diverse document types, and then fine-tuned or adjusted for specific scanning conditions. This partial training strategy reduces overall processing time by avoiding complete retraining for each document while maintaining high correction accuracy through condition-specific adaptations

Inventive Principle:

Principle #16Partial or excessive action

Data Source

PatentUS20250330547A1Image forming apparatus, image processing method, and non-transitory recording medium

Publication Date: 2025.10.23 RICOH CO LTD

AI summary

An image forming apparatus includes: a scanner to read a document to obtain first read data; and circuitry to acquire a training condition relating to the first read data, generate training data including the first read data and the training condition to be used for training processing based on machine learning, and perform show-through correction on second read data read by the scanner, using a trained model generated by the training processing using the training data.