Image Processing Device for Receipt Character Recognition Accuracy

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing character recognition technologies face challenges in accurately recognizing characters of varying sizes on documents, such as receipts, due to changes in aspect ratios, leading to incorrect recognition results.

Innovation Solution

An image processing device and method that extracts special character regions with different sizes and converts them to a basic size for improved recognition, using a combination of binarization, histogram analysis, and OCR engine processing to enhance character recognition accuracy.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If character recognition processing is performed on characters of varying sizes directly, then processing speed is maintained, but character recognition accuracy deteriorates for special size characters

Engineering Contradiction:
Improvecharacter recognition accuracyVSAvoidprocessing complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent segments characters into two categories: normal size characters and special size characters (enlarged or reduced). This segmentation allows the system to apply different processing strategies to different character types, improving recognition accuracy for special size characters while maintaining efficient processing for normal characters.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent changes the size parameter of special size characters by converting them to standard size through image processing operations (expansion or contraction). This parameter transformation enables the character recognition unit to process all characters uniformly, improving accuracy without significantly increasing processing complexity.

Inventive Principle:
Principle #35Parameter changes

2Measurement precision

If character size normalization is performed on all characters, then character recognition accuracy improves, but processing time increases

Engineering Contradiction:
Improvecharacter recognition accuracyVSAvoidprocessing time
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

Instead of applying size normalization to all characters (excessive action), the patent applies the operation only to special size characters (partial action). This selective approach improves recognition accuracy for problematic characters while avoiding unnecessary processing time consumption for normal characters.

Inventive Principle:
Principle #16Partial or excessive action

Solution Approach 2:

The patent performs preliminary identification and classification of characters into normal and special size categories before the main recognition process. This preliminary action allows the system to prepare special size characters for normalization in advance, optimizing the overall processing timeline.

Inventive Principle:
Principle #10Preliminary action

3Ease of manufacture

If no size conversion is performed, then processing simplicity is maintained, but recognition errors increase for enlarged or reduced characters

Engineering Contradiction:
Improveprocessing simplicityVSAvoidrecognition reliability
Core Design Contradiction:
Ease of manufactureVSReliability

Solution Approach 1:

The patent applies local quality by treating different character regions differently: normal size characters are processed directly while special size characters undergo size conversion. This localized approach maintains processing simplicity for the majority of characters while improving reliability for the minority of special size characters.

Inventive Principle:
Principle #3Local quality

Data Source

PatentEP3540644B1Image processing device, image processing method, and image processing program
Publication Date: 2021.09.22 SEIKO EPSON CORP
  • EP3540644B1 patent drawingFigure 1
  • EP3540644B1 patent drawingFigure 2
  • EP3540644B1 patent drawingFigure 3A

AI summary

An image processing device includes an acquisition unit that acquires a read image generated by reading a receipt or a bill and including a special character having a size different from a basic size of a character, a conversion unit that performs conversion processing on the size of the special character included in the read image into a size close to the basic size, and a character recognition unit that performs character recognition processing on the conversion-processed read image.