Hair Feature Transfer Using Segmented Generative Image Editing

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Current methods for editing hair characteristics in images often fail to produce photorealistic results and do not adequately preserve the identity of the person.

Innovation Solution

A computer-implemented method using a generative neural network to transfer hair features from a reference image to a digital image, combining face and background areas to create a composite modified image.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If current methods for editing hair characteristics are used, then the editing process can be performed, but the output images are not photorealistic and identity is not preserved

Engineering Contradiction:

Improvephotorealism and identity preservationVSAvoidhair feature transfer accuracy

Core Design Contradiction:

ReliabilityVSManufacturing precision

Solution Approach 1:

The image is divided into three distinct regions: face area, hair area, and background area. This segmentation allows independent processing of each region, enabling precise hair feature transfer while preserving the original face and background, thus resolving the contradiction between photorealism and editing precision

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

A generative neural network acts as an intermediary to transfer hair characteristics from a reference image to the target image. The network processes the hair region separately and reconstructs it with improved photorealism while maintaining identity, bridging the gap between current editing limitations and desired outcomes

Inventive Principle:

Principle #24Intermediary (Mediator)

2Reliability

If a generative neural network is used to transfer hair features, then photorealistic images with preserved identity can be generated, but the processing complexity increases

Engineering Contradiction:

Improvephotorealism and identity preservationVSAvoidgenerative neural network processing

Core Design Contradiction:

ReliabilityVSDevice complexity

Solution Approach 1:

By segmenting the image into face, hair, and background regions, the complex task of generating photorealistic hair while preserving identity is broken down into manageable steps. Only the hair region undergoes complex generative processing, while face and background remain unchanged, reducing overall processing complexity

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The generative neural network is applied selectively only to the hair region rather than the entire image. This local application maintains high photorealism where needed while minimizing unnecessary processing elsewhere, balancing quality improvement with computational efficiency

Inventive Principle:

Principle #3Local quality

Data Source

PatentUS12586272B2Methods and systems for transferring hair characteristics from a reference image to a digital image

Publication Date: 2026.03.24 ULTA SALON COSMETICS & FRAGRANCE

AI summary

An example computer-implemented image processing method includes receiving a digital image including a face area of a first subject and a background area, receiving a reference image of a second subject, modifying the digital image based on the reference image utilizing a generative neural network to create a modified digital image in which a feature of the second subject is transferred to the first subject, and combining the face area of the first subject and the background area from the digital image with the feature area from the modified digital image to create a composite modified digital image that includes data representing the feature of the second subject transferred to the first subject from the modified digital image and data representing the face area of the first subject and the background area from the digital image.