Technological development facilitated to fulfill the need to get a high-quality welding. The industries such as oil,aerospace and other important industries rely on reliable welding operations. Whereas the Radiographic testing is a well-established non-destructive testing method to detect subsurface welding defects. Aim of this research is to implement an automatic computer-aided identification system to recognize different types of welding defects in radiographic images which includes defect detection and classification. Deep neural network is used to build an efficient framework for weld defect classification. While on the other hand unavailability of standard preprocessed datasets, very limited data available in literature, imbalanced classification, overfitting during the training process and computational efficiency are the challenges which need to be encountered in this research.
https://ieeexplore.ieee.org/abstract/document/9467263
