A Computer Assisted Detection Framework of Kidney Diseases Based on CNN Model

Abstract

Kidney Tumors (KT) are the tenth most prevalent tumor in both men and women globally. Using image processing techniques, we provide several approaches and models in this research to identify renal illnesses from the dataset. In building our model, we employed deep learning CNN (Convolutional Neural Network), Keras, VGG16, SVM, U-NET, and Water-Shed. As an illustration, we found 100% accuracy in the Keras model and 50% accuracy in the VGG16 model. To develop these models, we initially only employed four classes: normal, cyst, stone, and tumor. Evaluating our model against the input photos gives accurate results to determine what stage of kidney illness it is in. We have tried our model multiple times to ensure it can reliably identify the condition from a single photograph. If our model is refined, it may be used to help the medical community diagnose kidney illnesses and administer the appropriate care.

Ähnliche Arbeiten