Face Deblurring via Aligned Grid Matching

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing face deblurring methods in image processing technology suffer from poor deblurring effects due to inaccurate detection of blurred regions and subsequent restoration of clear images, particularly in images captured with handheld devices that experience brightness imbalance and blurring.

Innovation Solution

A face deblurring method that involves aligning and dividing face images into grids, matching each grid with a dictionary of blurred images obtained through three-dimensional reconstruction, and querying corresponding clear grids from another dictionary to generate a clear image, effectively addressing posture variations and improving deblurring accuracy.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If blurred region detection and restoration algorithms are used on blurred images, then deblurring processing can be performed, but the deblurring effect is poor due to inaccurate detection of blurred regions

Engineering Contradiction:
Improveblurred region detection accuracyVSAvoiddeblurring effect quality
Core Design Contradiction:
Measurement precisionVSReliability

Solution Approach 1:

The patent performs alignment and grid division on the face image before deblurring processing. By pre-aligning the face image to a standard pose and dividing it into grids, the system prepares the image structure in advance, enabling more accurate matching with dictionary grids and improving subsequent deblurring accuracy.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent transforms the deblurring problem from direct pixel-level processing to grid-level processing in a transformed coordinate system. By changing the representation parameter from raw pixels to aligned grids, the system achieves more robust feature matching and improves detection accuracy of blurred regions.

Inventive Principle:
Principle #35Parameter changes

2Adaptability or versatility

If face images with different postures are processed using traditional deblurring methods, then processing can be performed, but the deblurring effect deteriorates due to posture variations

Engineering Contradiction:
Improveposture variation handlingVSAvoiddeblurring effect consistency
Core Design Contradiction:
Adaptability or versatilityVSReliability

Solution Approach 1:

The patent divides the face image into multiple grids after alignment, and processes each grid independently by matching with corresponding grids from the dictionary. This segmentation allows the system to handle local variations in different face regions, improving adaptability to posture changes while maintaining consistent deblurring quality across the entire face.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent uses a pre-built dictionary containing aligned face images with different postures. By copying and matching grids from this dictionary, the system can find appropriate reference patterns for various postures, enabling consistent deblurring performance across different facial orientations without requiring retraining.

Inventive Principle:
Principle #26Copying

Data Source

PatentEP3598385B1Face deblurring method and device
Publication Date: 2021.06.16 SUZHOU KEDA TECH
  • EP3598385B1 patent drawingFigure 1~3
  • EP3598385B1 patent drawingFigure 4
  • EP3598385B1 patent drawingFigure 5

AI summary

Disclosed in the present invention are a face deblurring method and device. The method comprises: acquiring a face image to be processed; aligning the face image to be processed onto a face mask, and performing grid division on the same; matching each grid of the divided face image to be processed with a grid of a first grid dictionary, so as to obtain a plurality of blurred grids corresponding to each grid of the face image to be processed, wherein the first grid dictionary is obtained by dividing a first two-dimensional image library according to the face mask after alignment; according to the blurred grids, querying in a second grid dictionary a plurality of clear grids corresponding to the plurality of blurred grids on a one to one basis, wherein the second grid dictionary is obtained after dividing a second two-dimensional image library according to the face mask alignment, and the blurred images correspond to the clear images on a one to one basis; and according to the queried clear grids, generating a clear image of the face image to be processed. The face deblurring method of an embodiment of the present invention can process face images of different postures and has an improved face deblurring effect.