Convex Clustering for Adaptive Chromatic Content Modeling

Solution Overview

Problem

Existing methods for modeling chromatic objects, such as images and color palettes, face challenges in setting the complexity of models, leading to unnatural color transfer due to inappropriate settings of discrete or continuous models, which can result in artifacts.

Innovation Solution

A method that optimizes a convex objective function with weighted kernels in a perceptual color space to automatically determine the complexity of the model, using convex clustering to identify the number of weighted kernels, allowing for a mixture model with automatically set complexity.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If a fixed number of Gaussian functions is used in the mixture model, then the model complexity is controlled, but the color transfer appears unnatural and creates artifacts due to inappropriate model complexity for different images

Engineering Contradiction:

Improvecolor transfer qualityVSAvoidmodel complexity

Core Design Contradiction:

ReliabilityVSDevice complexity

Solution Approach 1:

The patent applies dynamics by making the model complexity adaptive rather than fixed. The number of Gaussian functions is automatically determined based on the chromatic content of each image through convex clustering, allowing the model complexity to dynamically adjust to the specific requirements of different images. This resolves the contradiction by enabling appropriate model complexity for each image while maintaining reliable color transfer quality.

Inventive Principle:

Principle #15Dynamics

Solution Approach 2:

The patent changes the parameter of model complexity from a fixed value to a dynamically determined value. By using convex clustering to automatically count the number of significant Gaussian functions needed for each image, the system adapts the model complexity parameter to match the actual chromatic content requirements, eliminating artifacts while avoiding over-simplification.

Inventive Principle:

Principle #35Parameter changes

2Measurement precision

If the model complexity is increased to capture all color variations, then the representation accuracy improves, but the memory storage and retrieval time increase

Engineering Contradiction:

Improvechromatic content representation accuracyVSAvoidretrieval time

Core Design Contradiction:

Measurement precisionVSLoss of time

Solution Approach 1:

The system performs self-service by automatically determining the optimal model complexity for each image through convex clustering. The algorithm self-adjusts the number of Gaussian functions based on the image's chromatic content, selecting only the necessary number of functions to accurately represent the color distribution. This eliminates the need for manual complexity setting and optimizes both accuracy and retrieval performance.

Inventive Principle:

Principle #25Self-service

Solution Approach 2:

The patent applies parameter changes by dynamically adjusting the model complexity parameter for each image based on its chromatic content. The convex clustering algorithm automatically counts the number of significant Gaussian functions needed, changing the model complexity parameter adaptively. This ensures high representation accuracy for images with diverse colors while using minimal complexity for images with limited color variation, thereby reducing retrieval time.

Inventive Principle:

Principle #35Parameter changes

3Productivity

If a simple model is used for all images, then the processing speed is maintained, but the model cannot adapt to images with widely differing color distributions

Engineering Contradiction:

Improveprocessing speedVSAvoidadaptability to different color distributions

Core Design Contradiction:

ProductivityVSAdaptability or versatility

Solution Approach 1:

The patent applies dynamics by making the model complexity adaptive rather than fixed. The number of Gaussian functions is automatically determined based on the chromatic content of each image through convex clustering, allowing the model complexity to dynamically adjust to the specific requirements of different images. This resolves the contradiction by enabling appropriate model complexity for each image while maintaining reliable color transfer quality.

Inventive Principle:

Principle #15Dynamics

Solution Approach 2:

The patent changes the parameter of model complexity from a fixed value to a dynamically determined value. By using convex clustering to automatically count the number of significant Gaussian functions needed for each image, the system adapts the model complexity parameter to match the actual chromatic content requirements, eliminating artifacts while avoiding over-simplification.

Inventive Principle:

Principle #35Parameter changes

Data Source

PatentUS8379974B2Convex clustering for chromatic content modeling

Publication Date: 2013.02.19 GENESEE VALLEY INNOVATIONS LLC

AI summary

A system and method are provided for modeling a chromatic object, such as an image. For a set of colors of a chromatic object that are expressed as color values in a perceptual color space, the method includes optimizing a convex objective function which is a log likelihood function of a combination of weighted kernels centered on each color in the set over each of the other colors in the set. A number Nc of weighted kernels in the optimized function which each have a weight which is at least greater than 0 is identified. The chromatic object is modeled with a mixture model in which the complexity of the model is based on the identified number Nc.