Application of deep learning for the detection of default in fabric texture

Abstract

In terms of quality control, manual inspection of the fabric is time-consuming and inefficient. In this work, we are studying several models of deep convolutional neural networks (DCNNs) to prospect for fabric and detect manufacturing defects from real-time images. DCNNs have a powerful feature extraction and feature fusion capability that can simulate learning in the human brain. In order to improve computational efficiency and detection accuracy, the learning process consists of several convolution operations and the image features are extracted and processed step by step. Experimental results show that the best performance is obtained by the Detectnet model.