Deep Learning Clothing Style Evaluation System

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing clothing style evaluation methods via telecommunication networks require human intervention, leading to subjective and inconsistent results, and are time-consuming and costly, with a lack of objective and immediate feedback.

Innovation Solution

A deep learning-based method and system that analyzes images to extract fashion features, including skin tone, clothing type, color, and style, providing objective and immediate evaluation by calculating a total score using a predetermined equation, and recommending clothing based on tone-on-tone and tone-in-tone matching.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If human intervention is used for clothing style evaluation, then evaluation can be conducted with human judgment and feedback, but evaluation time increases and consistency and objectivity deteriorate

Engineering Contradiction:

Improveevaluation objectivityVSAvoidevaluation time

Core Design Contradiction:

Measurement precisionVSLoss of time

Solution Approach 1:

The patent replaces the mechanical system of human visual inspection and judgment with an automated image processing system using computer vision technology. The system automatically detects clothing items, extracts features, and evaluates styling without human intervention, thereby eliminating time loss while maintaining evaluation objectivity through consistent algorithmic processing.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The evaluation system performs self-assessment by automatically analyzing uploaded images, extracting fashion features, and generating evaluation results without requiring human evaluators. The system serves itself by implementing the complete evaluation workflow through automated algorithms, eliminating dependency on human resources while ensuring consistent and objective results.

Inventive Principle:

Principle #25Self-service

2Measurement precision

If human intervention is used for clothing style evaluation, then expert feedback can be obtained, but labor costs increase

Engineering Contradiction:

Improveevaluation reliabilityVSAvoidpersonnel expenses

Core Design Contradiction:

Measurement precisionVSQuantity of substance

Solution Approach 1:

The patent substitutes human experts with an automated computer vision system that performs clothing style evaluation. This replacement eliminates personnel expenses while maintaining evaluation reliability through consistent application of evaluation algorithms and criteria, removing the need for paid human evaluators.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The system creates a digital copy of the evaluation process by implementing automated algorithms that replicate human evaluation criteria and standards. This digital copying allows the system to perform evaluations reliably without requiring actual human experts, thereby eliminating labor costs while preserving evaluation quality.

Inventive Principle:

Principle #26Copying

3Adaptability or versatility

If virtual objects are used for clothing coordination, then virtual try-on and coordination can be achieved, but physical features of the actual user are not considered

Engineering Contradiction:

Improveclothing coordination capabilityVSAvoiduser feature accuracy

Core Design Contradiction:

Adaptability or versatilityVSMeasurement precision

Solution Approach 1:

The patent segments the user's actual physical appearance by detecting and extracting specific features such as skin tone, body shape, and physical characteristics from uploaded images. This segmentation allows the system to separate and analyze individual user attributes, enabling accurate matching with clothing items based on real physical features rather than generic virtual avatars.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The system changes the evaluation parameters from generic virtual object attributes to specific user physical feature parameters. By detecting and utilizing parameters such as skin tone, body measurements, and physical characteristics, the system adapts the clothing coordination capability to match the user's actual physical features, improving both adaptability and measurement precision.

Inventive Principle:

Principle #35Parameter changes

Data Source

PatentUS10915744B2Method for evaluating fashion style using deep learning technology and system therefor

Publication Date: 2021.02.09 FEELINGKI CO LTD

AI summary

Deep learning-based clothing style evaluation method and system are provided. According to the present invention, a deep learning-based clothing style evaluation method comprises receiving at least one image provided by a user terminal, analyzing objects included in the at least one image and extracting a plurality of fashion features related to a user using deep learning algorithm when the user exists in the object, extracting matching information between the extracted plurality of fashion features using deep learning algorithm, evaluating a clothing style of the user based on the matching information, and transmitting an evaluation result to the user terminal, wherein the fashion feature is one of the number of clothing worn by the user, type of clothing, color of clothing, style of clothing, skin tone of the user, and other information affecting fashion style, and wherein the matching information comprises essentially of matching information between the skin tone of the user and any one of the fashion features other than the skin tone.