Image Recolorization for Reducing Distractor Saliency

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Conventional systems struggle to remove distractor objects from images without distorting the image or making it look unrealistic, especially when maintaining the structure of the object is crucial.

Innovation Solution

A machine-trained model manipulates the chromatic information of distractor objects while preserving their luminance, using a palette transform generated by a machine-learned inpainting model to recolorize the distractors and harmonize them with the background, thus reducing saliency.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Object-affected harmful factors

If the distractor object is removed from the image, then the distraction is reduced, but the image becomes distorted or unrealistic

Engineering Contradiction:
Improvevisual distractionVSAvoidimage realism
Core Design Contradiction:
Object-affected harmful factorsVSReliability

Solution Approach 1:

The patent applies color changes by manipulating the chromatic information of the distractor object to match the surrounding background. The system identifies the distractor region and adjusts its color properties (hue, saturation, brightness) to harmonize with adjacent pixels, making the object visually blend into the background while preserving its structural presence in the image.

Inventive Principle:
Principle #32Color changes

Solution Approach 2:

The patent employs parameter changes by modifying specific chromatic parameters of the distractor object. The system extracts color parameters from the distractor region and transforms them to match the statistical distribution of colors in the surrounding background area, thereby reducing the object's visual salience without removing it entirely.

Inventive Principle:
Principle #35Parameter changes

2Object-affected harmful factors

If the distractor object is removed from the image, then the distraction is reduced, but the image structure is distorted

Engineering Contradiction:
Improvevisual distractionVSAvoidimage structure
Core Design Contradiction:
Object-affected harmful factorsVSShape

Solution Approach 1:

The patent applies the extraction principle by isolating the chromatic information of the distractor object from its luminance and structural information. The system selectively removes only the color properties that cause distraction while preserving the object's shape, position, and brightness, allowing the structure to remain intact while the distracting visual properties are extracted and replaced.

Inventive Principle:
Principle #2Taking out (Extraction)

3Object-affected harmful factors

If the distractor object is removed from the image, then the distraction is reduced, but additional processing complexity is required

Engineering Contradiction:
Improvevisual distractionVSAvoidprocessing complexity
Core Design Contradiction:
Object-affected harmful factorsVSDevice complexity

Solution Approach 1:

The patent applies self-service by enabling the system to automatically identify, analyze, and process distractor objects without requiring manual intervention. The system autonomously detects potential distractors, determines appropriate color transformations based on surrounding context, and applies the harmonization process, thereby reducing the need for complex manual editing tools or extensive user input.

Inventive Principle:
Principle #25Self-service

Data Source

PatentEP4156084B1Techniques for reducing distractions in an image
Publication Date: 2026.05.06 GOOGLE LLC
  • EP4156084B1 patent drawingFigure 1A
  • EP4156084B1 patent drawingFigure 1B
  • EP4156084B1 patent drawingFigure 1C

AI summary

Techniques for reducing a distractor object in a first image are presented herein. A system can access a mask and the first image. A distractor object in the first image can be inside a region of interest and can have a pixel with an original attribute. Additionally, the system can process, using a machine-learned inpainting model, the first image and the mask to generate an inpainted image. The pixel of the distractor object in the inpainted image can have an inpainted attribute in chromaticity channels. Moreover, the system can determine a palette transform based on a comparison of the first image and the inpainted image. The transform attribute can be different from the inpainted attribute. Furthermore, the system can process the first image to generate a recolorized image. The pixel in the recolorized image can have a recolorized attribute based on the transform attribute of the palette transform.