Art Image Sentiment Extraction Using Perceptual Color Analysis

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Current methods for extracting mood or sentiments from artwork images lack consideration of human subjective perception, fail to accurately identify dominant colors, and do not account for the unique characteristics of art images, leading to inefficiencies in personalized services and increased costs due to manual curation.

Innovation Solution

A method and system using an artificial neural network to preprocess art images, detect dominant perceptual colors, identify dominant subjects, and extract low-level features, classifying these into mood/sentiment classes to predict the mood or sentiment present in the image, incorporating human perception and reducing reliance on manual metadata generation.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If manual curation method is used to extract sentiments from artwork, then accuracy of mood identification may be improved, but cost and time consumption increase significantly

Engineering Contradiction:

Improvemood identification accuracyVSAvoidcuration time

Core Design Contradiction:

Measurement precisionVSLoss of time

Solution Approach 1:

The system enables automatic self-service mood extraction from artwork images using deep learning models. The neural network automatically analyzes color histograms, texture features, and compositional elements to identify sentiments without requiring manual curator intervention, thereby resolving the contradiction between accuracy and time consumption.

Inventive Principle:

Principle #25Self-service

Solution Approach 2:

The patent replaces the mechanical manual curation process with an automated electronic system based on computer vision and deep learning technologies. The system substitutes human curators with algorithms that process visual features and color information to extract sentiments automatically.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

2Ease of manufacture

If traditional color extraction methods are used, then processing simplicity is maintained, but human subjective perception is not accounted for leading to inaccurate sentiment extraction

Engineering Contradiction:

Improveprocessing simplicityVSAvoidcolor perception accuracy

Core Design Contradiction:

Ease of manufactureVSMeasurement precision

Solution Approach 1:

The system transforms color data from standard RGB values to perceptual color spaces that account for human subjective perception. By changing the parameter representation of colors and incorporating psychological color theory, the system maintains processing simplicity while significantly improving color perception accuracy for sentiment extraction.

Inventive Principle:

Principle #35Parameter changes

3Productivity

If pixel count based dominant color detection is used, then computational efficiency is improved, but accuracy in identifying perceptually dominant colors deteriorates

Engineering Contradiction:

Improvecolor detection speedVSAvoiddominant color accuracy

Core Design Contradiction:

ProductivityVSMeasurement precision

Solution Approach 1:

The system changes the parameters used for dominant color detection from simple pixel counts to perceptual weightings based on human color perception models. This allows the system to maintain computational efficiency while accurately identifying colors that are perceptually dominant rather than merely numerically frequent.

Inventive Principle:

Principle #35Parameter changes

4Productivity

If automated methods are implemented for sentiment extraction, then productivity and cost-effectiveness are improved, but reliability of sentiment classification may deteriorate without proper human guidance

Engineering Contradiction:

Improveautomation efficiencyVSAvoidsentiment classification reliability

Core Design Contradiction:

ProductivityVSReliability

Solution Approach 1:

The system incorporates feedback mechanisms where the deep learning model continuously learns from labeled artwork data and refines its sentiment classification capabilities. The automated process includes validation steps and iterative improvement through training on curated datasets, ensuring reliability while maintaining high productivity.

Inventive Principle:

Principle #23Feedback

Data Source

PatentUS20230022364A1Method and system for extracting sentiments or mood from art images

Publication Date: 2023.01.26 SAMSUNG ELECTRONICS CO LTD

AI summary

A method for extracting sentiments or mood from art images includes: receiving at least one of the art images as an input image; preprocessing the input image; extracting features from the preprocessed input image, the extracting including predicting a color label corresponding to a dominant perceptual color detected from the preprocessed input image a dominant subject from the preprocessed input image, detecting low-level image features from the preprocessed input image, and extracting mood feature information based on a description information included in the input image; classifying the extracted features into a plurality of mood/sentiments classes, using an artificial neural network; and predicting at least one of a mood or a sentiment that is present in the input image based on the dominant perceptual color and the plurality of mood/sentiments classes.