CAD System for Detection and Classification of Liver Cancer using Optimization Neural Network & Convolution Neural Network Classifiers
Reshma Jose, Shanty Chacko · 2020 International Conference on Power, Instrumentation, Control and Computing (PICC) · 2020
Deep learning is a recent field of machine learning that has garnered a great deal of attention in recent years. It has been commonly used for a variety of applications and has proved to be a strong machine learning method for many complex problems. Liver cancer appears to be a leading cause of female death, and a lot of money has been spent in the form of preventative screening programs. The use of automatic image processing techniques resulting from deep learning in that same sense represents a promising way of helping to detect liver cancer. In this paper, ADF-USM (Anisotropic diffusion filtering with Unsharp masking) completed the liver malignant growth CT (Computer Tomography) image preprocessing, the shapes and curves were effectively differentiated and upgraded by histogram equalization. It is seen by analyzing the results of the division method that the SSBIC (Superpixel Segmentation Based Iterative Clustering) algorithm produces favorable results than other current strategies. Finally, the process for classification relies mostly on execution of AGWO-CNN (Adaptive Grey Wolf Optimization with Convolution Neural Network classifier) to verify whether the images are benign or malignant.Experimental results on CT images show that the AGWOCNN model achieved high processing efficiency with an accuracy of 97.6 percent in the classification of liver cancer compared to other classification models.