Vol. 12, Special Issue 11 (2023)

Author(s):
Dan Mani Binu, Binu KM and Bindiya TS

Abstract:
This paper presents a dynamic comparative study on current machine learning (ML) applications and other state of the art techniques in the detection of renal calculi. Compared to other obstacles in the physiological system, the occurrence of kidney stones will not cause a high rate of mortality but may result in high morbidity around the world. Both Non-ML and ML techniques discussed in this paper are based on imaging techniques, viz. Ultrasound (US), magnetic resonance imaging (MRI), and computed tomography (CT) for detection of kidney stones. Multiple issues, viz. low-quality image, analysis of size over time, and kidney stone similarity are the hindering factors in the detection of renal calculi. In order to select the appropriate technique to make detection easier, a detailed analysis of several ML and non-ML algorithms has been carried out. Both techniques were found to have advantages and disadvantages. The decision for an effective treatment strategy for kidney stone detection was made easy by introducing the developing ML techniques, which still require additional improvements in advanced diagnostics, early detection, and innovative methodologies. A variety of scientific articles have been collected and reviewed to provide support for developing a real-time system. Perhaps it is very important to exploit the methods that provide accuracy in detecting kidney stones.