Kid-ML: ML For Kidney Malignant Tissues Identification

Document Type : Original Article

Authors

1 Electrical Communication and Electronics Systems Dept., School of Engineering, October University for Modern Sciences and Arts (MSA), Al Wahat Road, Giza, Egypt.

2 Systems and Information Department, Engineering and Renewable Energy Research Institute, National Research Centre (NRC), Dokki, Giza, Egypt. Postal Code: 12622.

Abstract

 A considerable worldwide medical and health burden is imposed by
kidney disease due to its high rates of morbidity and death as well as its high
economic cost. Imaging tests can be used by doctors to detect kidney tumors or
other diseases. Imaging studies include Magnetic Resonance Imaging(MRI),
Computed Tomography(CT) scan, and ultrasound scan which consume a lot of
time from doctors to detect kidney cancers through them. In order to help doctors
to identify tumors in their early stages, they can use simple Machine
Learning(ML) techniques or Deep Learning techniques through diagnostics and
predictions applications. A rise in interest in deep learning algorithms, which are
Artificially Intelligently (AI) based, on a worldwide scale has enabled recent
improvements in medical imaging and kidney segmentation. Deep Learning
techniques which are AI-based can offer and identify the kidney tumor in a more
efficient method, allowing for the development of a more effective kidney tumor
detection system. An input layer, one or more hidden layers, and an output layer  are the components of Artificial Neural Networks(ANNs) which is one kind of
Deep Learning(DL) algorithm that imitates biological neurons. Another kind is
Convolutional Neural Networks(CNNs) which are often the most effective and
well-liked in computer vision for image categorization in medical imaging. Deep
learning techniques based on CNNs have shown promising results in a variety of
medical image processing applications. However; all deep learning techniques
consume very high computational power. In this work, we study the use of simple
machine learning algorithms such as Decision Tree(DT), K Nearest
Neighbor(KNN), Random Forest(RF) and Logistic Regression(LR) algorithms
and compare their results. Simple machine learning algorithms consumes
minimum computational power. We discuss the behavior of those machine
learning algorithms while changing the resolution of the images. We found that
the accuracy of simple machine learning algorithm is stable while decreasing the
resolution of images to be 32pixels.

Keywords

Main Subjects

MSA Engineering Journal
Volume 2, Issue 2 - Serial Number 6
Special issue for Artificial Intelligence & Sustainable Engineering Conference: A Path to The Future
March 2023
Pages 1120-1134

Files

Share

How to cite

Statistics