Image Processing Apparatus Merging Binary and Edge Area Identification

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Conventional binarization methods require separate processing for reversed characters, leading to increased processing time and potential errors in character recognition, especially when dealing with complex document images.

Innovation Solution

An image processing apparatus that generates binary image data by performing binarization processing on input images, followed by edge extraction and area identification, merging overlapping character areas identified from both binary and edge images to reduce processing time and prevent erroneous detections, allowing for partial binarization of character areas.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If separate processing is performed for reversed characters using conventional binarization methods, then character recognition can be performed, but processing time increases and errors may occur

Engineering Contradiction:
Improvecharacter recognition accuracyVSAvoidprocessing time
Core Design Contradiction:
ReliabilityVSLoss of time

Solution Approach 1:

The patent merges the binarization results from the first binarization processing (for normal characters) and the second binarization processing (for reversed characters) into a single output image. The merging unit combines the character recognition results from both processing paths, allowing the system to handle both reversed and non-reversed characters in a unified manner, thereby reducing processing time while maintaining recognition accuracy.

Inventive Principle:
Principle #5Merging (Combining)

Solution Approach 2:

The patent dynamically selects which binarization result to use based on the actual character orientation detected in the image. The system performs both binarization processes and then determines whether characters are reversed or not, selecting the appropriate result or merging both results accordingly. This dynamic approach allows flexible adaptation to different document types without fixed processing paths.

Inventive Principle:
Principle #15Dynamics

2Measurement precision

If partial binarization processing is performed on character areas based on edge images and binary images separately, then character areas can be identified, but processing time increases due to redundant operations

Engineering Contradiction:
Improvecharacter area identification accuracyVSAvoidprocessing time
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The patent merges overlapping character areas identified from both edge image processing and binary image processing into a unified character area map. By identifying the intersection and union of character areas from both methods, the system eliminates redundant processing operations while maintaining comprehensive character area detection, thus reducing processing time without sacrificing identification accuracy.

Inventive Principle:
Principle #5Merging (Combining)

Data Source

PatentUS7751648B2Image processing apparatus, image processing method, and computer program
Publication Date: 2010.07.06 CANON KK
  • US7751648B2 patent drawing
  • US7751648B2 patent drawing
  • US7751648B2 patent drawing

AI summary

An image processing apparatus generates binary image data by performing binarization processing on input image data and generates edge image data by performing edge extraction processing to the input image data. The image processing apparatus acquires a first area identification result by performing area identification processing on the generated binary image data and acquires a second area identification result by performing area identification processing on the generated edge image data. In addition, the image processing apparatus generates a merged area identification result by merging the first area identification result and the second area identification result and generates an output binary image by performing partial binarization processing on the input image data according to each character area included in the generated merged area identification result.