Facial Image Wavelet Blending for Natural Skin Texture Restoration

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing face enhancement technologies remove underlying skin textures, resulting in artificial or fake-like skin in enhanced facial images.

Innovation Solution

A method and electronic device that utilize wavelet decomposition and weightage factors based on facial characteristics to generate a texture restored image, enhancing the source image while retaining natural skin textures and removing blemishes.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If traditional face recognition algorithms are used, then the processing speed is fast, but the recognition accuracy is low and cannot distinguish similar faces

Engineering Contradiction:
Improveface recognition accuracyVSAvoidprocessing system complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent divides the face image processing into multiple stages: initial face detection, landmark point identification, feature extraction, and recognition. By segmenting the complex recognition task into smaller sub-tasks, the system achieves higher accuracy without overwhelming computational complexity.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The system performs preliminary actions by first detecting face landmarks and extracting key features before performing the actual recognition. This preliminary processing prepares the data in advance, enabling more accurate recognition while maintaining efficient processing speed.

Inventive Principle:
Principle #10Preliminary action

2Measurement precision

If more feature points are extracted to improve recognition accuracy, then the recognition precision increases, but the processing time increases

Engineering Contradiction:
Improveface recognition precisionVSAvoidprocessing time
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

Instead of uniformly processing all face regions, the patent identifies and processes only the most discriminative local features such as landmark points and key facial contours. This selective processing of locally important regions maintains high recognition precision while reducing overall processing time.

Inventive Principle:
Principle #3Local quality

Solution Approach 2:

The system dynamically adjusts the number and type of features extracted based on the specific recognition task requirements. By changing parameters such as landmark point density and feature extraction depth, the system optimizes the balance between precision and processing speed for different scenarios.

Inventive Principle:
Principle #35Parameter changes

3Measurement precision

If the face recognition system processes high-resolution images, then the recognition accuracy improves, but the energy consumption increases

Engineering Contradiction:
Improverecognition accuracyVSAvoidenergy consumption
Core Design Contradiction:
Measurement precisionVSUse of energy by moving object

Solution Approach 1:

The patent extracts only the essential and most informative features from high-resolution face images, such as landmark points and key contour features, rather than processing the entire high-resolution image data. This extraction approach maintains recognition accuracy while significantly reducing the computational energy required.

Inventive Principle:
Principle #2Taking out (Extraction)

Solution Approach 2:

The system applies partial action by processing only the critical portions of the face image that contain discriminative information for recognition. By focusing computational resources on these partial regions rather than the entire image, the system achieves accurate recognition with lower energy consumption.

Inventive Principle:
Principle #16Partial or excessive action

Data Source

PatentEP3874456B1Method and electronic device for processing facial images
Publication Date: 2026.05.06 SAMSUNG ELECTRONICS CO LTD
  • EP3874456B1 patent drawingFigure 1
  • EP3874456B1 patent drawingFigure 2A
  • EP3874456B1 patent drawingFigure 2B

AI summary

A method includes determining information indicative of at least one facial characteristic associated with at least one face in the source image, processing the source image using a filter based on the determined information, performing wavelet decomposition on each of the filtered image and the source image, determining weightage factors associated with the wavelet decomposition of each of the filtered image and the source image, based on the determined information, obtaining a wavelet image to generate a texture restored image from the wavelet decomposition of each of the filtered image and the source image based on the weightage factors.