Abstract:Textile Research Journal, Ahead of Print. Color-patterned fabrics possess changeable patterns, low probability of defective samples, and various forms of defects. Therefore, the unsupervised inspection of color-patterned fabrics has gradually become a research hotspot in the field of fabric defect detection. However, due to the redundant information of skip connections in the network and the limitation of post-processing, the current reconstruction-based unsupervised fabric defect detection methods have difficulty in detecting some defects of color-patterned fabrics. In this article, we propose an Attention-Gate-based U-shaped Reconstruction Network (AGUR-Net) and a dual-threshold segmentation post-processing method. AGUR-Net consists of an encoder, an Atrous Spatial Pyramid Pooling module and an attention gate weighted fusion residual decoder. The encoder is used to obtain more representative features of the input image via EfficientNet-B2. The Atrous Spatial Pyramid Pooling module is used to enlarge the receptive field of the network and introduce multi-scale information into the decoder. The attention-gate-weighted residual fusion decoder is used to fuse the features of the encoder with the features of the decoder to obtain the reconstructed image. The dual-threshold segmentation post-processing is used to obtain the final defect detection results. Our method achieves a precision of 59.38%, a recall of 59.1%, an F1 of 54.31%, and an intersection-over-union ratio of 41.18% on the public dataset YDFID-1. The experimental results show that the proposed method can better detect and locate the defects of color-patterned fabrics compared with several other state-of-the-art unsupervised fabric defect detection methods.

Fabric image recolorization by fuzzy pretrained neural network

Recoloring Textile Fabric Images Based on Improved Fuzzy Clustering

Recolouring Digital Textile Printing Design with High Fidelity

An Unsupervised Method for Dominant Colour Region Segmentation in Yarn-Dyed Fabrics

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

CTH-Net: CNN-Transformer Hybrid Network for Garment Image Generation from Sketches and Color Points

Deep Exemplar-Based Color Transfer for 3D Model

Learning the Color Space for Effective Fabric Defect Detection

Learning Natural Colors for Image Recoloring.

SCSNet: an Efficient Paradigm for Learning Simultaneously Image Colorization and Super-resolution

Image Colorization Algorithm Based on Deep Learning

Color Segmentation of Multicolor Porous Printed Fabrics by Conjugating SOM and EDSC Clustering Algorithms

Real-Time User-Guided Image Colorization with Learned Deep Priors

Learning Spatially Decoupled Color Representations for Facial Image Colorization

Woven Fabric Pattern Recognition and Classification Based on Deep Convolutional Neural Networks

Attention-Gate-based U-shaped Reconstruction Network (AGUR-Net) for color-patterned fabric defect detection

Dyeing Creation: a Textile Pattern Discovery and Fabric Image Generation Method

Hyperspectral reconstruction from a single textile RGB image based on the generative adversarial network

Deep Palette-Based Color Decomposition For Image Recoloring With Aesthetic Suggestion

Real-time Neural Woven Fabric Rendering

ROMNet: Renovate the Old Memories