Face Region Color Correction Using Segmented Candidate Detection

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing display devices inaccurately correct facial skin color due to neighboring objects with similar colors being mistakenly altered, leading to errors in image processing.

Innovation Solution

A display device equipped with an image processor that detects a face candidate region using face detection algorithms and performs skin color correction, dividing the image into preset color regions to prioritize the correction of the facial region over overlapping colors, ensuring only the face is accurately colored.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If skin color correction is applied to regions with similar colors, then skin color correction effectiveness is improved, but neighboring objects with similar colors are mistakenly corrected

Engineering Contradiction:
Improveskin color correction effectivenessVSAvoidcorrection accuracy
Core Design Contradiction:
Measurement precisionVSReliability

Solution Approach 1:

The patent segments the image processing into distinct stages: first identifying face candidate regions using color information, then applying face detection algorithms to verify actual face regions, and finally performing skin color correction only on confirmed face regions. This segmentation prevents neighboring objects with similar colors from being mistakenly corrected by requiring multiple verification steps before applying correction.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent applies different processing strategies to different regions of the image. Face candidate regions are identified using color distribution, face detection algorithms are applied specifically to these candidates, and skin color correction is localized only to confirmed face regions. This local quality approach ensures that correction is applied with appropriate precision to each region based on its verified status.

Inventive Principle:
Principle #3Local quality

2Measurement precision

If face detection algorithm is applied to all regions, then face detection accuracy is improved, but processing time and complexity increase

Engineering Contradiction:
Improveface detection accuracyVSAvoidprocessing time
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The patent performs preliminary color-based segmentation to identify face candidate regions before applying the face detection algorithm. By pre-filtering the image to identify only regions with skin-like colors, the face detection algorithm needs to process only these candidate regions rather than the entire image, significantly reducing processing time while maintaining detection accuracy.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent applies the face detection algorithm partially, only to regions that have been pre-identified as face candidates based on color information. This partial action approach avoids the excessive processing that would result from applying the algorithm to the entire image, while still ensuring that all actual face regions are detected within the candidate set.

Inventive Principle:
Principle #16Partial or excessive action

Data Source

PatentEP2884733B1Display device and method for correcting colors of a face image
Publication Date: 2019.12.25 SAMSUNG ELECTRONICS CO LTD
  • EP2884733B1 patent drawingFigure 1
  • EP2884733B1 patent drawingFigure 2
  • EP2884733B1 patent drawingFigure 3

AI summary

A display device which accurately detects and corrects only a facial region contained in an image and thereby prevents an error in which a background or neighboring object having a similar color to a skin color is corrected together with the facial region, and a method of controlling the same are provided. The display device includes an image processor configured to detect a face candidate region from an input image, based on information associated with at least one face color region, to detect a face region by applying a face detection algorithm to the detected face candidate region and to perform skin color correction on the detected face region, and a display configured to display the corrected input content.