Pakistan Journal of Engineering & Technology (Sep 2023)
Graphical User Interface-Based Detection of Kidney Stones Using Image Segmentation Techniques
Abstract
The exponential increase in detrimental surroundings and unhealthy nourishments is causing various health issues in humans. The most destructive effect of such lifestyles is on the kidneys, which cause many kidney diseases, and the most common among them are kidney stones. Kidney stones are a regular but life-threatening disease as they mostly remain unrecognized at the initial stages, leading to an increased threat of end-stage kidney failure. Due to the high recurring rate, Medical Imaging technologies are of paramount importance in detecting this serious public health threat worldwide. This research paper provides a Graphical User Interface for the detection of kidney stones that enables ease of understanding and point-and-click control of the algorithm in MATLAB. The proposed work uses CT scan images to explore image processing techniques due to their reliability and regularity. The algorithm enhances kidney stone screening by improving image quality and focusing on the region of interest. This study highlights the best solutions of imaging techniques to resolve problems like grainy pixels, low-resolution images, and the inaccurate detection of kidney stones due to size and resemblance with nearby parts. All this begins with examining the medical imaging slices from the body area, which later undergoes preprocessing, segmentation, and boundary detection techniques. To check the accuracy of the algorithm, 12 features were extracted using GLCM. Finally, the obtained features were classified using the Classification Designer App in MATLAB. An accuracy of 99.2% was acquired using Ensemble Classifiers. Also, for further progress in the detection of kidney stones in the future, an AI model can be trained so that it can deal with images having different thresholds for better management of the disease.
Keywords
