Identification of Crystals Present in a Urine Sediment based on Adaptive Boosting Algorithm

Abstract

Urinalysis is one of the most common examination being done to check the components present in the urine. A microscopic exam is needed to detect certain components present in urine such as Red blood cells, White blood cells, and crystals. Certain diseases can be seen in the urine through the form of urine crystals. The objective of the study is to detect the urine crystal present in the patient’s urine by image processing after undergoing centrifugation of the urine sample. A microscope was used with a Raspberry Pi 2 mounted on it and a Raspberry Pi camera placed on the eyepiece of the microscope to capture the image of the urine sediment. The process used the application of Harr feature. Adaptive Boosting was used before sending the data to the support vector machine. This study is limited in detecting the urine crystal provided by the medical laboratories in the country. The study will be important in the health sector specifically in detecting urinary tract abnormalities by classifying the type of crystal present in the urine. 30 sample urine images were done by the researchers. The testing gathered an accuracy of 90% when compared to the traditional urinalysis.