Image Processing Apparatus Text Area Identification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current image processing techniques face challenges in accurately identifying and separating text areas from other content in images, especially when the background level is high, leading to inefficient compression and potential errors in text recognition.
Innovation Solution
An image processing apparatus and method that uses a processor to extract feature elements from image data, determine conditions based on these elements, and set areas and attributes within the image data, employing determination conditions to distinguish text areas from other content, such as dividing lines, background dividing-lines, and specific characters, to apply appropriate compression ratios and enhance text legibility.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional image processing techniques are used to identify text areas, then processing speed is maintained, but text area identification accuracy deteriorates in images with high background levels
Solution Approach 1:
The patent segments the image processing task into multiple stages: extracting candidate lines, calculating background levels for each line, evaluating text area probabilities, and iteratively refining selections. This multi-stage segmentation allows the system to handle high background levels by processing information in controlled steps rather than attempting global analysis, thereby improving text area identification accuracy despite background interference.
Solution Approach 2:
The patent applies local quality analysis by calculating background levels individually for each candidate line rather than using a global background assumption. Each line's text area probability is evaluated based on its local characteristics (foreground pixel density, background level, position relationships). This local adaptation enables accurate text area identification even when different regions have varying background levels, directly addressing the problem of high background level interference.
2Productivity
If uniform compression ratio is applied to entire image, then processing complexity is reduced, but compression efficiency deteriorates for documents with mixed text and image areas
Solution Approach 1:
The patent segments the image into distinct text areas and non-text areas based on evaluated probabilities and determination conditions. This segmentation enables differential compression where text areas use higher compression ratios and image areas use lower compression ratios, significantly improving overall compression efficiency for mixed-content documents while maintaining acceptable processing complexity through automated classification.
Solution Approach 2:
The patent assigns different compression qualities to different regions of the image based on their classified attributes. Text areas receive aggressive compression with higher ratios, while image areas receive conservative compression with lower ratios. This local quality differentiation directly improves compression efficiency for mixed-content documents without requiring manual intervention, balancing productivity gains with automated complexity management.
3Loss of substance
If aggressive compression is applied to entire image, then file size is reduced, but text recognition accuracy deteriorates
Solution Approach 1:
The patent applies different compression qualities to different regions: text areas use aggressive compression with higher ratios to achieve significant file size reduction, while image areas use conservative compression with lower ratios to preserve visual quality and text recognition accuracy. This local quality differentiation ensures that text regions maintain sufficient fidelity for OCR while achieving overall file size reduction, directly resolving the contradiction between compression aggressiveness and text recognition reliability.
Data Source
AI summary
An image processing apparatus includes a memory and at least one processor or circuitry or a combination thereof. The memory is configured to store determination conditions. The at least one processor or circuitry or the combination thereof is configured to acquire image data and extract a feature element from the image data. The at least one processor or circuitry or the combination thereof is further configured to determine whether the extracted feature element satisfies conditions defined in the determination conditions. The at least one processor or circuitry or the combination thereof is further configured to set an area with reference to an extracted feature element satisfying the conditions based on the determination conditions. The at least one processor or circuitry or the combination thereof is further configured to set an attribute of the area based on the determination conditions.