Kidney Tumor Segmentation and Classification using Deep Neural Network on CT Images

Kidney disease is one of many severe chronic diseases that a person can have. Early detection of this disease can be pivotal for proper treatment. Different neural networks have proven to be useful in disease prediction in the progression of modern science. In this paper, we have proposed a segmentation based kidney tumor classification method using deep neural network (DNN). We have done our work in two steps. Firstly, we have segmented kidneys using a manual segmentation technique utilizing UNet along with SegNet. Then, for the classification task, the MobileNetV2, VGG16, and InceptionV3 models are trained on the segmented kidney data. “CT KIDNEY DATASET: Normal-Cyst-Tumor and Stone” dataset (published in Kaggle) has been used to train our models. Finally, the classification models MobileNetV2, VGG16, InceptionV3 scored 95.29%, 99.48% and 97.38% accuracy on test set. We have found that the VGG16 model has the best accuracy with the highest sensitivity and specificity. Explainable AI (GradCam) method has been applied to explain our model's result.