Hierarchical Document Rendering for Accurate Character Recognition
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing OCR methods face decreased accuracy when recognizing characters in full-page image documents or outline-font-character documents, especially when characters are represented in intermediate colors or located on complex images, as they require conversion to full-page images for processing.
Innovation Solution
An information processing apparatus and method that analyzes documents represented by hierarchical drawing commands, determines if commands draw graphic objects, renders these objects into images, and performs character recognition on the rendered images, rather than converting the entire document into a full-page image.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the entire document is converted to a full-page image for OCR processing, then the document can be processed uniformly, but the character recognition accuracy decreases when characters are represented in intermediate colors or located on complex images
Solution Approach 1:
The patent segments the document into individual graphic objects by analyzing drawing commands at different hierarchical levels. Instead of processing the entire document as one image, it extracts and processes each graphic object separately, which improves character recognition accuracy while reducing the complexity of handling complex documents.
Solution Approach 2:
The patent extracts graphic objects from the document hierarchy by identifying drawing commands that create graphic objects. These extracted objects are then rendered and processed independently by the OCR system, separating the character recognition task from the complex background context.
2Adaptability or versatility
If outline-font-character documents are converted to full-page images, then OCR processing can be performed, but the character recognition accuracy decreases due to loss of text object information
Solution Approach 1:
The patent performs preliminary analysis of the document hierarchy and drawing commands before OCR processing. It identifies and extracts graphic objects in advance, rendering them separately so that the subsequent OCR processing can focus on clear, isolated character images rather than complex full-page images.
Solution Approach 2:
The patent creates a rendered image copy of each extracted graphic object for OCR processing. This copy preserves the character information while eliminating the complexity of the original document structure, allowing accurate recognition without needing to process the entire original document as an image.
3Stability of the object's composition
If characters are represented in intermediate colors or located on other images, then the document structure is preserved, but the character recognition accuracy decreases in conventional OCR methods
Solution Approach 1:
The patent applies different processing qualities to different parts of the document. By analyzing drawing commands at specific hierarchical levels, it identifies graphic objects with local characteristics (such as intermediate colors or positions on other images) and processes them individually, maintaining their structural information while improving recognition accuracy.
Data Source
AI summary
An information processing apparatus input a document represented by hierarchical structure including a plurality of drawing commands; analyze each of the plurality of drawing commands of respective hierarchical levels corresponding to from a rearmost side to a frontmost side in the document, starting from the rearmost side; determine whether or not the drawing command being analyzed is a command for drawing a graphic object; generate a rendered image of the graphic object by rendering the graphic object in a case where the determining determines that the drawing command is a command to draw the graphic object; and perform character recognition processing on the rendered image.