Urine Sediment Analysis by Using Convolution Neural Network

Abstract

Urinary particles are important requirements in clinical urinalysis, particularly in the diagnosis and monitoring of patients suspected of renal diseases and urinary tract Infections. As a result, it is critical to identify urinary particles accurately in the clinical area. Also the outcome is hugely affected by the doctor's experience. However, because the traditional manual microscopic analysis relies on human operators who read the samples visually and identify them, this method is slow, time-consuming, and labor-intensive, In this research, presented a deep learning method for analyzing urinary particles. The authors prepare a dataset of urine sediment microscopic images, which includes approximately 820 cell annotations and four-cell classes: RBC, Calcium oxalate, cysteine calcium, and uric acid. Used for deep learning training and testing of various convolutional network models. The authors proposed Convolution neural network structure and five ConvNet models such as MobileNet, VGG16, DenseNet, ResNet50V, InceptionV3. According to these evaluations, the best models for true positive recall are MobileNet, and the proposed method ist he second one.. These models also achieve the highest accuracy of 98.3 percent. on the other hand, InceptionV3 and DenceNet have comparable accuracy results with 96.5 percent.