Skin Tone Pixel Processing for HDR Image Detail Preservation

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Image processing techniques, especially in high dynamic range (HDR) processing, often result in unwanted artefacts and loss of detail in skin tone regions, leading to an unrealistic appearance in images.

Innovation Solution

A method that detects skin tone pixels, applies local contrast enhancement and color adjustment, followed by local sharpening using filters like anti-aliasing and sharpener filters, to regenerate the image, thereby preserving and enhancing details in skin areas without affecting non-skin tone regions.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Manufacturing precision

If local sharpening is applied to skin tone pixels, then detail preservation in skin areas is improved, but processing complexity increases

Engineering Contradiction:

Improvedetail preservationVSAvoidprocessing complexity

Core Design Contradiction:

Manufacturing precisionVSDevice complexity

Solution Approach 1:

The image processing is segmented into distinct stages: skin tone detection, local contrast enhancement, color adjustment, and local sharpening. Each stage processes only skin tone pixels identified in the detection phase, avoiding unnecessary processing of non-skin areas and reducing overall computational complexity while maintaining detail preservation.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

Different processing operations (contrast enhancement, color adjustment, sharpening) are applied specifically to skin tone regions rather than the entire image. This localized approach preserves skin details while minimizing processing requirements for non-skin areas, resolving the contradiction between detail preservation and processing complexity.

Inventive Principle:

Principle #3Local quality

2Reliability

If multiple processing steps (contrast enhancement, color adjustment, sharpening) are applied to skin tone pixels, then image realism is improved, but processing time increases

Engineering Contradiction:

Improveimage realismVSAvoidprocessing time

Core Design Contradiction:

ReliabilityVSLoss of time

Solution Approach 1:

Skin tone pixels are detected and identified in advance before the main processing steps are applied. This preliminary detection allows subsequent contrast enhancement, color adjustment, and sharpening operations to be applied only to relevant pixels, reducing total processing time while maintaining image realism through multiple processing steps.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The processing pipeline applies contrast enhancement, color adjustment, and sharpening in a continuous sequence to skin tone pixels without interrupting the processing flow. This continuous action on identified skin regions maintains realism while optimizing processing efficiency by avoiding repeated detection and selection steps.

Inventive Principle:

Principle #20Continuity of useful action

Data Source

PatentEP3460747B1Method, processing system and computer program for processing an image

Publication Date: 2023.09.06 VESTEL ELEKTRONIK SANAYI & TICARET ANONIM SIRKETI

Figure 1
Figure 2~3

AI summary

In a method of processing an image formed of pixels, skin tone pixels in the image are detected (20). Local sharpening is applied (50) to at least some of the skin tone pixels. The image is regenerated (60) using the locally sharpened skin tone pixels.