Image Recognition Apparatus Preventing Skew Correction Information Loss

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

When recognizing characters on an input image, existing techniques often result in lost information due to skew correction, as only the aligned recognition area is displayed, hiding important information outside this area, which can be uncomfortable for users and affect recognition accuracy.

Innovation Solution

An information processing apparatus that aligns an input image with a template image using a transformation matrix, generates a check image including both the template and aligned input images, and displays this alongside the recognition result, ensuring that the correspondence between the images and the result is recognizable, thereby preventing information loss and enhancing user comfort.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Ease of operation

If the input image is corrected for skew and then displayed, then the user comfort is improved, but information outside the recognition area might be lost

Engineering Contradiction:
Improveuser comfortVSAvoidinformation outside recognition area
Core Design Contradiction:
Ease of operationVSLoss of information

Solution Approach 1:

The patent merges the corrected input image and the template image into a single composite image for display. The composite image includes both the recognition area (from template) and the full corrected input image, allowing users to view both aligned recognition results and complete image information simultaneously, thus preventing information loss while maintaining alignment benefits

Inventive Principle:
Principle #5Merging (Combining)

Solution Approach 2:

The patent transitions from displaying only the recognition area (2D cropped view) to displaying a composite image that incorporates both the recognition area and the full corrected image in a unified coordinate system. This dimensional expansion allows preservation of all image information while maintaining the aligned recognition area for comparison

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

2Manufacturing precision

If only the recognition area of the template image is displayed, then the alignment between template and input image is achieved, but information outside the recognition area is lost

Engineering Contradiction:
Improvealignment precisionVSAvoidinformation outside recognition area
Core Design Contradiction:
Manufacturing precisionVSLoss of information

Solution Approach 1:

The patent combines the template image and the full corrected input image into a composite display structure. The template defines the recognition area boundaries, while the composite image preserves all original image information outside this area, allowing both precise alignment and complete information display

Inventive Principle:
Principle #5Merging (Combining)

Solution Approach 2:

The patent segments the display into distinct functional areas: the recognition area (defined by template boundaries) and the surrounding areas (containing additional image information). This segmentation allows the system to maintain precise alignment in the recognition area while preserving and displaying information in external areas through the composite image structure

Inventive Principle:
Principle #1Segmentation

Data Source

PatentUS11258925B2Information processing apparatus for displaying the correction of an image and non-transitory computer readable medium
Publication Date: 2022.02.22 FUJIFILM BUSINESS INNOVATION CORP
  • US11258925B2 patent drawing
  • US11258925B2 patent drawing
  • US11258925B2 patent drawing

AI summary

An information processing apparatus includes a receiver that receives an input image to be recognized and a processor configured to, by executing a program, align the input image with a template image in such a way as to match a recognition area of the input image and a recognition area defined on the template image, perform a process for recognizing the recognition area of the aligned input image, generate a check image including the template image and the aligned input image, and display the check image and a result of the process for recognizing the recognition area of the aligned input image such that a correspondence between the check image and the result is recognizable.