A Morphological Image-based Recognition of Iron Triad using a Convolutional Neural Network

초록

A processed image typically results in numerous imperfections in the binary regions due to thresholding. By considering the shape and organization of the image, morphological image processing aims to overcome these shortcomings. In this research, the proposed focus is to develop a neural network model that can automatically identify and categorize elements of the Iron Triad using a Morphological Image-based Recognition through a Convolutional Neural Network (CNN). This tool extracts image components to represent and describe specific image characteristics. This paper describes the convolution-based classification neural network's architecture, which extract the critical features of the image dataset. It shows that the plot of the convolutional neural network training progress had achieved a better validation accuracy. Experiments show that the method presented is effective, providing a classification accuracy of 96.5%, which is much better than other research works that use CNN.

키워드