Image Processing Apparatus Dynamic Masking for Recognition Accuracy
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image processing technologies face challenges in accurately generating image data that does not represent information of a deletion target, as they often rely on fixed processing methods regardless of the information present in the image data, leading to suboptimal recognition accuracy.
Innovation Solution
An information processing apparatus equipped with a processor that adapts cleansing processing based on specific information within the image data, such as keywords or document types, to generate image data that excludes the deletion target information while preserving relevant information, using cleansing learning devices and character recognition learning devices implemented with artificial intelligence.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If fixed processing is applied to generate image data excluding deletion target information, then processing simplicity is maintained, but recognition accuracy deteriorates
Solution Approach 1:
The patent implements dynamic image processing that adapts to different document types and information contents. The processor determines the document type (e.g., family register, residence certificate) and selectively applies different processing methods to generate appropriate masked images, rather than using a fixed processing approach. This dynamic adaptation resolves the contradiction by maintaining processing simplicity through automation while significantly improving recognition accuracy through content-aware processing.
Solution Approach 2:
The patent changes processing parameters based on the detected document type and information content. Different masking strategies are applied depending on the specific document characteristics - for example, different approaches for family registers versus residence certificates. This parameter adaptation allows the system to maintain operational simplicity while achieving high recognition accuracy by optimizing processing for each document type.
2Loss of time
If fixed masking processing is used regardless of document content, then processing time is reduced, but image data quality deteriorates
Solution Approach 1:
The patent performs preliminary document type determination and content analysis before applying masking processing. By pre-identifying the document type (family register, residence certificate, etc.) and understanding the information structure, the system can select the appropriate processing method in advance. This preliminary action enables efficient processing without sacrificing image data quality, as the correct masking strategy is applied from the outset rather than requiring iterative adjustments.
Solution Approach 2:
The system dynamically adjusts processing methods based on real-time document analysis. The processor adapts the masking approach according to the specific document type and content structure detected, optimizing both processing efficiency and output quality. This dynamic processing resolves the contradiction by automatically selecting the most appropriate processing path for each document, maintaining speed while ensuring quality.
3Device complexity
If generic image processing is applied to all document types, then device complexity is reduced, but processing effectiveness deteriorates
Solution Approach 1:
The patent segments the image processing task into distinct stages: document type determination, content analysis, and selective masking application. Each stage handles specific aspects of the processing, allowing the system to manage complexity through modular design while maintaining high effectiveness. The segmentation enables specialized processing for different document types without requiring a completely different system for each type.
Solution Approach 2:
The patent implements a universal processing framework that handles multiple document types (family registers, residence certificates, etc.) through a single integrated system. The processor determines document type and automatically adapts the processing approach, providing multi-functional capability without requiring separate dedicated systems for each document type. This universality maintains reasonable system complexity while ensuring effective processing across diverse document types.
Data Source
AI summary
An information processing apparatus includes a processor. The processor is configured to receive first image data, and generate, by processing corresponding to information represented in the first image data and corresponding to specific information other than information of a deletion target out of the information represented in the first image data, second image data not representing the information of the deletion target out of the information represented in the first image data but representing the information other than the information of the deletion target.