Clothing Style Identification via Image Segmentation and Feature Detection

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing clothing recommendation systems fail to provide personalized and style-based suggestions, as they do not accurately distinguish between different clothing items in an image and do not consider demographic data, leading to inadequate and cumbersome shopping experiences for individuals.

Innovation Solution

A computing device with a segment module to identify regions of related clothing within an image, a detection module to analyze features, and a processor to determine clothing style, combined with a recognition module to collect demographic data, enabling personalized and style-based recommendations.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Extent of automation

If clothing recommendations are based on shopping history from a specific merchant, then the recommendations are automated and consistent with past purchases, but the recommendations are not personalized to the individual's style and may not be applicable to different users

Engineering Contradiction:

Improveautomation of clothing recommendationsVSAvoidpersonalization to individual style

Core Design Contradiction:

Extent of automationVSAdaptability or versatility

Solution Approach 1:

The system segments the clothing recommendation process into multiple independent modules: a segment module that divides an image into regions of different clothing items, a detection module that identifies features of each item, and a processor that analyzes features to determine style. This segmentation enables the system to handle multiple clothing items independently and generate personalized recommendations based on each item's style characteristics.

Inventive Principle:

Principle #1Segmentation

2Productivity

If clothing recommendations are based on similarity to analyzed clothing, then the recommendations are generated automatically, but the system does not distinguish between different clothing items and does not consider the person's style

Engineering Contradiction:

Improvespeed of generating recommendationsVSAvoidaccuracy of style identification

Core Design Contradiction:

ProductivityVSMeasurement precision

Solution Approach 1:

The segment module partitions an image containing multiple clothing items into distinct regions, allowing the system to analyze each item separately. This enables precise identification of style features for each clothing item while maintaining efficient automated processing.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The system introduces an intermediary analysis process between image capture and recommendation generation. The detection module extracts style features from segmented clothing regions, and the processor analyzes these features to determine style characteristics, serving as an intermediary that bridges raw image data and personalized recommendations.

Inventive Principle:

Principle #24Intermediary (Mediator)

3Quantity of substance

If the region of interest captures multiple clothing items, then the system can analyze available clothing, but the recommendations become cumbersome and not style-based

Engineering Contradiction:

Improvenumber of clothing items analyzedVSAvoidusability of recommendations

Core Design Contradiction:

Quantity of substanceVSEase of operation

Solution Approach 1:

The segment module automatically divides an image containing multiple clothing items into distinct regions, with each region corresponding to a specific clothing item. This segmentation enables the system to process and analyze multiple items efficiently while maintaining clear organization, making the recommendations easier to navigate and use.

Inventive Principle:

Principle #1Segmentation

Data Source

PatentUS8861866B2Identifying a style of clothing based on an ascertained feature

Publication Date: 2014.10.14 HEWLETT PACKARD DEVELOPMENT COMPANY LP

AI summary

Examples disclose a method executed on a computing device to locate a clothing region within an image to segment into a region of related clothing. Further, the examples provide detecting a feature of the related clothing. Additionally, the examples also disclose determining a style of the related clothing in the region based on the detection of the feature.