Super-Resolution Method Using Local Binary Pattern Classification

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Conventional resolution enhancement techniques are time-consuming when converting Full-HD content to UHD quality, failing to meet the demand for rapid image quality improvement in display devices like 4K-UHD TVs.

Innovation Solution

A super-resolution method using local binary pattern classification and linear mapping through a mapping kernel to quickly convert low-resolution input images into high-resolution output images by separating input images into unit patches, classifying texture types, and generating high-resolution patches based on determined mapping kernels.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Manufacturing precision

If conventional resolution enhancement techniques are used to convert Full-HD content to UHD quality, then image quality is improved, but conversion time is excessively long

Engineering Contradiction:

Improveimage qualityVSAvoidconversion time

Core Design Contradiction:

Manufacturing precisionVSLoss of time

Solution Approach 1:

The input image is divided into multiple unit patches (e.g., 3x3 pixel blocks), and each patch is processed independently through LBP classification and mapping kernel application. This segmentation enables parallel processing of different image regions, dramatically reducing overall conversion time while maintaining UHD quality standards for each patch.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent transforms the resolution enhancement problem from a continuous pixel interpolation task into a discrete classification and mapping task. By converting pixel values to LBP patterns (binary representations) and applying pre-computed mapping kernels, the method changes the processing parameters from continuous intensity values to discrete categorical operations, enabling faster real-time processing.

Inventive Principle:

Principle #35Parameter changes

2Productivity

If real-time super-resolution is achieved through LBP classification and linear mapping, then conversion speed is improved by 2,650 times, but processing complexity increases

Engineering Contradiction:

Improveconversion speedVSAvoidprocessing complexity

Core Design Contradiction:

ProductivityVSDevice complexity

Solution Approach 1:

Mapping kernels are pre-computed and stored for each possible LBP pattern type before actual image processing. During real-time conversion, the system only needs to classify the current patch's LBP pattern and retrieve the corresponding pre-computed kernel, avoiding complex calculations during processing. This preliminary preparation dramatically reduces real-time computational complexity.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The patent replaces traditional mechanical/image processing methods (such as iterative optimization algorithms or complex interpolation techniques) with a lookup-based system. Instead of performing heavy computational mechanics during real-time processing, the system substitutes these with simple pattern matching and matrix multiplication operations using pre-computed kernels, achieving massive speedup.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Data Source

PatentUS11455707B2Super-resolution method using local binary pattern classification and linear mapping

Publication Date: 2022.09.27 UNIVERSITY INDUSTRY COOPERATION GROUP OF KYUNG HEE UNIVERSITY

AI summary

A super-resolution method according to the inventive concept may include receiving an input image of low resolution, separating the input image into low resolution (LR) unit patches, classifying a texture type of each of pixels included in the input image using a local binary pattern for the LR unit patches, generating high resolution (HR) unit patches corresponding to each of the pixels based on a mapping kernel corresponding to the texture type, and combining the HR unit patches based on a predetermined setting to generate an output image of high resolution.