Defect-Classification-from-Weld-Radiography-Images-Using-VGG-19-based-Convolutional-Neural-Network

The classification of welding defects in radiographic images is done by a CNN model which is trained by fine-tuning VGG-19 network.

GRIMA database of X-ray images (GDXray) is used in this work for preparing the dataset. The GDXray welding subset contains of 78 X-ray images of different weld specimens. https://domingomery.ing.puc.cl/material/gdxray/