Fabric defect classification using transfer learning and deep learning

The internal inspection of fabrics is one the most important phases production in order to achieve high quality standard textile industry. Therefore, developing efficient automatic control mechanism has been an extremely major area research. In this paper, famous architecture Googlenet was fine-tuned into two configurations for texture defect classification that trained on a database (TILDA). experimental result, both configurations, achieved significant overall accuracy score 97% motif and non-motif-based images 89% mixed images. results obtained, it observed second model, which updates last six layers, more successful than first one; layers.