Abstract:A novel optimisation technique based on the differential evolution (DE) algorithm with dynamic parameter selection (DPS-DE) is proposed to develop a colour difference classification model for dyed fabrics, improve the classification accuracy, and optimise the output regularisation extreme learning machine (RELM). The technique proposed is known as DPS-DE-RELM and has three major differences compared with DE-ELM: (1) Considering that the traditional ELM provides an illness solution based on the output weights, DE is proposed to optimise the output of the RELM. (2) Considering the simple parameter setting of the traditional algorithm, the DE algorithm with DPS is adopted. (3) For DPS, an optimal range of parameters is chosen, and the efficiency of the algorithm is significantly improved. This study analyses the colour difference classification of fabric images captured under standard lighting based on the DPS-DE-RELM algorithm. First, the colour difference of the fabric images is calculated and six color-difference-related features extracted, and second the features are classified into five different levels based on the perception of humans. Finally, a colour difference classification model is built based on the DPS-DE-RELM algorithm, and then the optimal classification model suitable for this study is selected. The experimental results show that the output method with regularisation parameters can achieve a maximum classification accuracy of 98.87%, which is higher compared with the aforementioned optimised original ELM algorithm, which can achieve a maximum accuracy of 84.67%. Therefore, the method proposed has the advantages of greater convergence speed, high classification accuracy, and robustness.

A superpixel‐based color difference detection algorithm for dyed fabrics

Recoloring Textile Fabric Images Based on Improved Fuzzy Clustering

An Efficient Method for Solid‐colour and Multicolour Region Segmentation in Real Yarn‐dyed Fabric Images

A Novel Method for Weft and Warp Yarn Segmentation in Multicolour Yarn‐dyed Fabric Images

Recolouring Digital Textile Printing Design with High Fidelity

Colour Matching Comparison Between Spectrophotometric and Multispectral Imaging Measurements

Fabric Image Registration Based on Affine Transform and Levenberg-Marquardt Algorithm

A Multispectral Imaging Approach to Colour Measurement and Colour Matching of Single Yarns Without Winding

Learning the Color Space for Effective Fabric Defect Detection

New Algorithm for Calculating Perceived Colour Difference of Images

Image Retrieval of Wool Fabric. Part II: Based on Low-Level Color Features

Optimization of Full Spectrum Color Matching Algorithm for Color Spun Yarn Based on Visual Characteristics

An Efficient Recoloring Method for Color Vision Deficiency Based on Color Confidence and Difference

A Research of Image Difference Algorithm Based on the Correlation of Color Attributes.

Perceptually-based color difference for complex images

Fabric Defect Detection Using Salience Metric for Color Dissimilarity and Positional Aggregation

A Color Image Edge Detection Algorithm Based on Color Difference

Color detection of printing based on improved superpixel segmentation algorithm

Colour Difference Classification for Dyed Fabrics Based on Differential Evolution with Dynamic Parameter Selection to Optimise the Output Regularisation Extreme Learning Machine

Color calibration for fabric image analysis based on spectral reflectance reconstruction

Image recoloring of printed fabric based on the salient map and local color transfer