Image Conversion Using Word Area Detection and Non-Linear Scaling

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Image scaling between different aspect ratios, such as 4:3 and 16:9, often results in distortion or loss of important content areas like words, as existing methods like linear scaling or cropping fail to retain and display them accurately.

Innovation Solution

An image conversion method that performs word area detection using a combination of edge-based and texture-energy-based algorithms, sets weight values for detected word and non-word areas, and applies a non-linear scaling algorithm to maintain aspect ratio while minimizing distortion, ensuring important content is retained and displayed clearly.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Ease of manufacture

If linear scaling algorithm is adopted for image conversion from 4:3 to 16:9, then the conversion process is simple, but the word area in the image is greatly distorted

Engineering Contradiction:
Improvesimplicity of conversion processVSAvoiddistortion of word area
Core Design Contradiction:
Ease of manufactureVSManufacturing precision

Solution Approach 1:

The patent applies different scaling algorithms to different regions of the image based on content type. Word areas are identified and processed with algorithms that preserve text clarity (such as nearest-neighbor or bicubic interpolation), while non-word areas use different scaling methods. This local differentiation resolves the contradiction by maintaining word area quality without requiring complete complexity throughout the entire image processing pipeline.

Inventive Principle:
Principle #3Local quality

Solution Approach 2:

The image is segmented into word areas and non-word areas through text detection algorithms. By dividing the image into these distinct regions, the patent can apply appropriate scaling strategies to each segment - preserving text integrity in word areas while allowing more flexible scaling in non-critical regions, thus resolving the distortion issue without uniform complexity.

Inventive Principle:
Principle #1Segmentation

2Manufacturing precision

If cropping algorithm is adopted to perform cropping on edges of an image, then image distortion is not caused, but all or a part of a word area of the image is likely to be lost

Engineering Contradiction:
Improveavoidance of image distortionVSAvoidloss of word area
Core Design Contradiction:
Manufacturing precisionVSLoss of information

Solution Approach 1:

The patent performs text detection and word area identification before the scaling operation. By detecting and marking word areas in advance, the system can preserve these regions during subsequent scaling operations, preventing both distortion and loss of text content. This preliminary action ensures that important information is protected before any transformation occurs.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent dynamically adjusts scaling parameters based on the detected content type. For word areas, scaling parameters are modified to prevent distortion (using higher quality interpolation methods), while non-word areas may use different parameters. This parameter adaptation resolves the contradiction by preventing distortion in critical areas without requiring uniform cropping that would lose information.

Inventive Principle:
Principle #35Parameter changes

3Manufacturing precision

If non-linear scaling algorithm with weight values is adopted for image conversion, then the word area is retained and displayed without distortion, but the conversion process becomes more complex

Engineering Contradiction:
Improveretention and clarity of word areaVSAvoidcomplexity of conversion process
Core Design Contradiction:
Manufacturing precisionVSDevice complexity

Solution Approach 1:

The patent applies weight values and different scaling algorithms only to specific word areas rather than the entire image. By localizing the complex processing to only where text is detected, the system achieves high precision in preserving word areas while minimizing the overall computational complexity burden, as non-word areas can be processed more simply.

Inventive Principle:
Principle #3Local quality

Data Source

PatentEP2426638B1Image conversion method, conversion device and display system
Publication Date: 2013.09.18 HUAWEI DEVICE CO LTD
  • EP2426638B1 patent drawingFigure 1~2
  • EP2426638B1 patent drawingFigure 3~4
  • EP2426638B1 patent drawingFigure 5

AI summary

An image conversion method, a conversion device, and a display system are provided in the embodiments of the present invention. The image conversion method includes: performing word area detection on an image to acquire a detected word area; and performing conversion processing on the image according to the word area to acquire a converted image that has an aspect ratio different from that of an unconverted image. The conversion device includes: a detection unit, configured to perform word area detection on an image to acquire a detected word area; and a conversion unit, configured to perform conversion processing on the image according to the word area to acquire a converted image that has an aspect ratio different from that of an unconverted image. In this way, an important content area of the image may be retained and clearly displayed.