Convolutional neural network based ResNet50 for finding accuracy in prediction of lung cancer using CT images and compared with CNN based inception V3

Abstract

To find and compare the accuracy in prediction of lung cancer using ResNet-50 and Inception V3. Materials and methods: The CT images with two classes (diseased/normal) are taken. The dataset related to our work is collected from the Iraq-Oncology Teaching Hospital/National Center for Cancer Diseases. Convolution Neural Network (CNN) based ResNet50 and Inception V3 methods are the two groups with sample size of 989 in.each with Gpower (80%). Results: The proposed model ResNet50 produced improved accuracy of (0.93094±0.057554%) than Inception-V3(0.69938±0.138030) with the significance value of <0.05. Conclusion: ResNet50 produced high accuracy (%) results compared with CNN-Inceptionv3 model.