CNN Image Processing for Flicker-Free Video Super-Resolution

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Flickering occurs in high-resolution images generated from low-resolution images due to inconsistent patterns in frames, and existing methods for motion estimation and compensation are complex and ineffective.

Innovation Solution

An image processing apparatus using a convolutional neural network to generate high-resolution images by extracting feature information from low-resolution images and removing flickering by utilizing previous frames' high-resolution images, thereby reinforcing frame correlation and skipping complicated motion estimation.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If interframe motion estimation and compensation is used to remove flickering, then flickering removal effect is improved, but processing complexity increases and correct motion estimation becomes difficult

Engineering Contradiction:
Improveflickering removal effectVSAvoidprocessing complexity
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The patent extracts only the necessary temporal correlation information from previous frames through simple subtraction operations, rather than performing complete motion estimation and compensation. This extracts the essential flickering removal capability while avoiding the complexity of full motion analysis

Inventive Principle:
Principle #2Taking out (Extraction)

Solution Approach 2:

Instead of estimating motion to compensate for flickering, the patent inverts the approach by directly subtracting previous frame information from current frame to eliminate flickering patterns, achieving flicker removal without traditional motion estimation

Inventive Principle:
Principle #13The other way round (Inversion)

2Productivity

If high-resolution image of current frame is not used in next frame generation, then processing speed is improved, but correlation between frames is weakened

Engineering Contradiction:
Improveprocessing speedVSAvoidframe correlation
Core Design Contradiction:
ProductivityVSReliability

Solution Approach 1:

The patent performs preliminary subtraction of previous frame information from the current frame to remove flickering before generating the final high-resolution image. This preliminary action ensures temporal correlation is established early, improving both frame correlation and subsequent processing efficiency

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent maintains continuous temporal correlation by using the generated high-resolution image of the current frame (after flicker removal) as input for the next frame's flicker removal process, creating a continuous chain of useful information flow across frames

Inventive Principle:
Principle #20Continuity of useful action

3Reliability

If complete high-resolution images are stored for flicker removal, then flickering removal accuracy is improved, but memory usage increases

Engineering Contradiction:
Improveflickering removal accuracyVSAvoidmemory usage
Core Design Contradiction:
ReliabilityVSQuantity of substance

Solution Approach 1:

The patent extracts only the essential feature information from low-resolution images rather than storing complete high-resolution images. This extraction provides sufficient data for accurate flicker removal while dramatically reducing memory requirements

Inventive Principle:
Principle #2Taking out (Extraction)

Solution Approach 2:

The patent uses lightweight feature representations instead of heavy full-resolution images, creating a more efficient, resource-light solution that achieves the same flicker removal functionality with minimal memory overhead

Inventive Principle:
Principle #27Cheap short-living objects (Disposable)

Data Source

PatentEP3790272B1Image processing apparatus and operation method thereof
Publication Date: 2025.10.01 SAMSUNG ELECTRONICS CO LTD
  • EP3790272B1 patent drawingFigure 1
  • EP3790272B1 patent drawingFigure 2
  • EP3790272B1 patent drawingFigure 3

AI summary

Provided is an image processing apparatus for generating a high-resolution image. The image processing apparatus includes a memory storing one or more instructions and a processor configured to execute the one or more instructions stored in the memory, wherein the processor is configured to extract feature information regarding a low-resolution image of a current frame by using a first convolutional neural network, generate, based on the feature information, a first high-resolution image of the current frame, remove flickering of the first high-resolution image by using a high-resolution image of a previous frame, and remove flickering of a high-resolution image of a next frame by using at least one of a flickering-removed second high-resolution image of the current frame, or the feature information.