Skin Hue Correction in Video Processing for Biometric Detection

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing methods for detecting biometric information from video data, such as pulse-based biometrics, are influenced by the magnitude of hue values and are not robust against variations in lighting and skin color, leading to inaccuracies in biometric information detection.

Innovation Solution

A video processing apparatus and method that corrects the hue values of skin areas in video frames to optimal values for biometric detection using correction coefficients, enhancing the detection of biometric information like pulse rate by aligning hue values to 50° and 230°, and optionally testing multiple corrections to find the best fit for individual video conditions.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If hue-based biometric detection is performed without correction, then the detection process is simple, but the detection accuracy is low due to influence from lighting and skin color variations

Engineering Contradiction:
Improvebiometric information detection accuracyVSAvoidcolor correction process complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent changes the hue parameter of skin area pixels by applying correction coefficients to transform the hue values to target hue values. This parameter transformation compensates for variations in lighting and skin color, thereby improving biometric information detection accuracy without requiring complex hardware modifications

Inventive Principle:
Principle #35Parameter changes

Solution Approach 2:

The patent introduces correction coefficients as an intermediary element between the original video data and the corrected video data. These coefficients, calculated based on skin area color information, serve as a mediator to adjust hue values and eliminate the influence of lighting and skin color variations on biometric detection

Inventive Principle:
Principle #24Intermediary (Mediator)

2Reliability

If correction coefficients are calculated and applied to skin area pixels, then the robustness against lighting and skin color variations is improved, but the processing time and computational complexity increase

Engineering Contradiction:
Improverobustness against lighting and skin color variationsVSAvoidvideo processing time
Core Design Contradiction:
ReliabilityVSLoss of time

Solution Approach 1:

The patent performs preliminary calculation of correction coefficients based on skin area color information before applying them to correct the video data. By preparing the correction factors in advance, the actual correction process becomes more efficient and can be applied uniformly across the video frames, reducing overall processing time

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent applies color correction specifically to the skin area pixels rather than the entire video frame. By identifying skin areas and applying correction only to those regions, the processing load is significantly reduced while maintaining robustness against lighting and skin color variations where it is most needed

Inventive Principle:
Principle #3Local quality

Data Source

PatentEP4163883B1Video processing apparatus, control method therefor, and program
Publication Date: 2025.11.05 CANON KK
  • EP4163883B1 patent drawingFigure 1~2
  • EP4163883B1 patent drawingFigure 3
  • EP4163883B1 patent drawingFigure 4

AI summary

This disclosure provides a video processing apparatus that detects biometric information of an object from a video of the object composed of a plurality of time-series images, wherein the apparatus comprises acquiring means for receiving the video input thereto, and acquiring a skin area of an object; determining means for determining correction coefficients based on a color of the skin area and a color indicating a target hue, the correction coefficients being for approximating a hue of pixels in the skin area to the target hue; correcting means for correcting, with use of the correction coefficients, the pixels in the skin area in each of the frames composing the video; and detecting means for detecting biometric information of the object based on time-series images of the skin area after the correction by the correcting means.