Foreground Segmentation for Image Similarity

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing image matching technologies in augmented reality applications face challenges in accurately determining visual similarity between images due to background interference, as they fail to effectively isolate and compare the foreground objects across different images.

Innovation Solution

A method and system that segregate the foreground from the background in an image, replace the background with solid or textured images, and generate combined histograms to compare the similarity between images, thereby isolating the characteristics of the foreground and reducing background interference.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Ease of manufacture

If histograms are calculated for entire images including background, then the comparison process is simple, but the accuracy of object similarity determination deteriorates due to background interference

Engineering Contradiction:

Improvesimplicity of comparison processVSAvoidaccuracy of object similarity determination

Core Design Contradiction:

Ease of manufactureVSMeasurement precision

Solution Approach 1:

The image is divided into foreground and background portions. The foreground portion containing the object of interest is segmented from the background, allowing separate processing. This enables accurate comparison of object regions while excluding interfering background elements from the histogram calculation.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The background portion is extracted and replaced with a uniform color or removed entirely from the image before histogram calculation. This extraction isolates the object of interest, ensuring that the histogram reflects only the object's visual characteristics without background contamination.

Inventive Principle:

Principle #2Taking out (Extraction)

2Measurement precision

If the background is replaced with different images to isolate foreground, then the accuracy of foreground matching is improved, but the complexity of the processing system increases

Engineering Contradiction:

Improveaccuracy of foreground matchingVSAvoidcomplexity of processing system

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

Different portions of the image are treated differently: the foreground portion is preserved with its original visual characteristics, while the background portion is replaced with a uniform color or simplified representation. This local differentiation improves matching accuracy without requiring complex processing of the entire image.

Inventive Principle:

Principle #3Local quality

Solution Approach 2:

The background replacement transforms the background portion from a complex visual scene into a uniform color or simplified pattern. This parameter change (from varied to uniform) reduces the background's contribution to the histogram, isolating the object characteristics while maintaining a relatively simple processing approach.

Inventive Principle:

Principle #35Parameter changes

Data Source

PatentUS9076076B1Image similarity determination

Publication Date: 2015.07.07 GOOGLE LLC

AI summary

A system and method is provided that determines whether objects in one image are visually similar to objects in another image by replacing the images' backgrounds with other images, such as a solid color or an image with texture, and comparing the resulting histograms.