Performance Analysis of Weld Image Classification Using Modified Resnet Cnn Architecture Using Modified Resnet Cnn Architecture, a Performance Analysis of Weld Image Classification

Abstract

The detection and classification of weld images is critical for improving the quality of joined
materials during the production process. To automate the classification of weld images in
industry, this paper proposes an effective automatic method for the detection and
classification of weld images into four different cases using deep learning methods. In this
work, a Convolutional Neural Network (CNN) is used for weld image classification by
modifying the CNN architecture's internal architecture. This proposed ResNet CNN
architecture consists of three Convolutional layers, two pooling layers with activation layers,
and two Fully Connected Neural Networks (FCNN). The FCNN in the proposed CNN
architecture is designed with 15 internal hidden layers, and each hidden layer is designed
with20 neurons which obtains high classification efficiency. The morphological activity
functional methods are used on the classified weld images to detect the crack regions.