Fabric Identification via Pixel Shift Histograms

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Current technologies lack the ability to accurately specify the fabric of clothing in images, particularly in online shopping contexts where numerous products require precise information, and existing methods are inefficient in terms of arithmetic operation load.

Innovation Solution

An image processing device that acquires and processes images by shifting them pixel by pixel to generate differential images, creating histograms of hue, saturation, and brightness components to determine the fabric based on unique reflection patterns.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If fabric specification technology is added to image processing, then product information accuracy is improved, but arithmetic operation load increases

Engineering Contradiction:

Improvefabric specification accuracyVSAvoidarithmetic operation load

Core Design Contradiction:

Measurement precisionVSUse of energy by moving object

Solution Approach 1:

The patent extracts only the essential features needed for fabric identification by calculating differences between shifted images and generating histograms of differential pixel values. This extraction approach focuses computational resources on the most discriminative features rather than processing the entire image data, thereby reducing arithmetic operation load while maintaining fabric specification accuracy

Inventive Principle:

Principle #2Taking out (Extraction)

Solution Approach 2:

The patent transforms the image data into a different parameter space by computing differential images through pixel shifting and then generating histograms. This parameter transformation converts complex fabric texture patterns into simplified statistical distributions that are easier and less computationally intensive to analyze for fabric identification

Inventive Principle:

Principle #35Parameter changes

2Measurement precision

If manual input for each product is required, then information accuracy is improved, but productivity decreases

Engineering Contradiction:

Improveproduct information accuracyVSAvoidproduct processing speed

Core Design Contradiction:

Measurement precisionVSProductivity

Solution Approach 1:

The patent implements self-service by enabling the system to automatically specify fabric types from product images without requiring manual input from operators. The automated fabric specification technology processes images and determines fabric characteristics independently, eliminating the need for time-consuming manual data entry while maintaining high accuracy, thus significantly improving productivity

Inventive Principle:

Principle #25Self-service

Data Source

PatentUS10867378B2Image processing device, image processing method, server, and computer program

Publication Date: 2020.12.15 RAKUTEN GROUP INC

AI summary

An image processing device acquires an image to be processed that includes an image of clothing, the image being composed of a hue, saturation and brightness; calculates a difference between the image to be processed and a plurality of shift images obtained by shifting the image to be processed in units of one pixel by at least one pixel or more in a prescribed direction, and generates a plurality of difference images that correspond to the number of shifted pixels; generates a histogram of differential pixel values for each of the hue, saturation and brightness of the difference images; determines the fabric of the clothing on the basis of the features of the histogram.