Fabric Defect Detection in Textile

Papers:

• Self-attention Deep Saliency Network for Fabric Defect Detection_CBCT_2020 [Paper]

• MLMA-Net: multi-level multi-attentional learning for multi-label object detection in textile defect images_arXiv_2021 [Paper]

• Fabric Defect Segmentation Method Based on Deep Learning_2020 [Paper]

• A Robust Fabric Defect Detection Method Based on Improved RefineDet_2020 [Paper]

• Automatic Fabric Defect Detection with a Multi-Scale Convolutional Denoising Autoencoder Network Model_2018

• Defective texture classification using optimized neural network structure_2020

• Research on Fabric Defect Detection Based on Deep Fusion DenseNet-SSD Network_2020

• Fabric Defect Detection System Using Stacked Convolutional Denoising Auto-Encoders Trained with Synthetic Defect Data_2020

• Yarn-dyed Fabric Defect Detection with YOLOV2 Based on Deep Convolution Neural Networks_2018

• Mobile-Unet: An efficient convolutional neural network for fabric defect detection_2020

• Fabric defect detection based on a deep convolutional neural network using a two-stage strategy_2020

• Automatic fabric defect detection using a wide-and-light network_2021

• Fabric Defect Detection Based on Cascade Faster R-CNN_2020

• Exploring Faster RCNN for Fabric Defect Detection_2020

• Automatic fabric defect detection with a wide-and-compact network_2018

• Automatic Fabric Defect Detection Method Using PRAN-Net_2020

• D4Net: De-deformation defect detection network for non-rigid products with large patterns_2020

• EDDs: A series of Efficient Defect Detectors for fabric quality inspection_2020

Self-attention Deep Saliency Network for Fabric Defect Detection_CBCT_2020 [Paper]

Motivation:

• Fabric defect detection is one of the key steps in the textile manufacturing industry. 

• Traditional saliency detection models mostly rely on hand-crafted features to obtain local details and global context. 

• However, these methods ignore the association between context features. It restricts the ability to detect salient objects in complex scenes. 

Objectives:

• A deep saliency detection model: incorporates self-attention mechanism into convolutional neural network.

  1. First, a fully convolutional network is designed for multi-scale feature maps to capture rich context features of the fabric image. 

  2. Then, after the side output of the backbone network, the self-attention module is adopted to coordinate the dependencies between the features of the multiple layers, which improves the characterization ability of the extracted features. 

  3. Finally, the multi-level saliency maps output from the self-attention mechanism are fused by the short connection structure and generating detail enriched saliency map. 

• Experiments demonstrate that the proposed method outperforms the state-of-the-art approaches when the defects are blurred or the shape is complex.

Category: Self-attention, Hybrid transformer, Saliency maps.

MLMA-Net: multi-level multi-attentional learning for multi-label object detection in textile defect images_arXiv_2021 [Paper]

Motivation:

• For the sake of recognizing and classifying textile defects, deep learning-based methods have been proposed and achieved remarkable success in single-label textile images. 

• However, detecting multi-label defects in a textile image remains challenging due to the co-existence of multiple defects and small-size defects. 

Objectives:

• A multi-level, multi-attentional deep learning network (MLMA-Net):

  1. Increase the feature representation ability to detect small-size defects; 

  2. Generate discriminative representation that maximizes the capability of attending the defect status, which leverages higher-resolution feature maps for multiple defects. 

  3. A multi-label object detection dataset (DHU-ML1000) in textile defect images. 

• The results demonstrate that the network extracts more distinctive features and has better performance than the state-of-the-art approaches on the real-world industrial dataset. 

Category: Multi-level, Multi-attentional, Dataset (DHU-ML1000)

Fabric Defect Segmentation Method Based on Deep Learning [Paper]

Motivation:

• Fabric defect detection plays an essential role in the textile production process, which was widely applied in the textile industry. 

• However, lots of important problems, such as the accuracy of detection, the computational complexity of the algorithm, and data imbalance, still needed to be addressed for application in industrial production. 

Objectives:

• An efficient convolutional neural network for defect segmentation and detection. 

  1. Alleviates the manual annotation cost of the data set

  2. Only needs few defect samples combined with standard samples to learn the potential feature of defects and obtain the location of defects with high accuracy. 

• The network is divided into two parts: segmentation and decision. 

  1. First, the fabric data set without training is utilized as the input of the segmentation network. 

  2. Then, the output of the segmentation network is applied as the raw materials for training the decision network. 

  3. Finally, a well-trained network is used to obtain the location of defects with high accuracy. 

• The proposed method only demands almost 50 defect samples to get accurate segmentation results and can achieve the requirement of real-time detection with a speed of 25 frames per second (FPS). 

Category: Fabric defect detection.

A Robust Fabric Defect Detection Method Based on Improved RefineDet_2020 [Paper]

Motivation:

• Modifying RefineDet for fabric defect detection.

Objectives:

• A robust fabric defect detection method, based on the improved RefineDet. 

  1. Firstly, the method uses RefineDet as the base model.

  2. Secondly, we design an improved head structure based on the Full Convolutional Channel Attention (FCCA) block and the Bottom-up Path Augmentation Transfer Connection Block (BA-TCB).

  3. Finally, the proposed method applies many general optimization methods, such as attention mechanism, DIoU-NMS, and cosine annealing scheduler, and verifies the effectiveness of these optimization methods in the fabric defect localization task. 

• Experimental results show that the proposed method is suitable for the defect detection of fabric images with unpattern background, regular patterns, and irregular patterns.

Category: Fabric defect detection, Attention, RefineDet.