Endoscopic Bladder Tissue Classification Using Seventeen Layered Deep Convolutional Neural Network

After a diagnosis of bladder cancer, endoscopic surgery is carried out as it is thought to be the most efficient and appropriate method of treating superficial tumors. It is essential to classify the bladder tissues accurately using MRI since early diagnosis of the problem can aid in treatment and perhaps save lives. Early-stage diagnosis is still challenging due to the delayed onset of indications and tiny alterations in urinary tract tissue that are imperceptible to the naked eye for diagnosis. The automated approach might be utilized to automatically determine the type of endoscopic bladder tissue by using the proper deep learning model with the necessary convolutional layers. This paper suggests the Seventeen Layered Deep Convolutional Neural Network (17-DCNN) to classify the bladder tissue classes effectively. The Endoscopic bladder tissue Dataset from Kaggle is used for the implementation of DOMN that contains 1785 images with four classes as HGC, LGC, NTL, and NST. The 17-DCNN model's initial phase is an analysis of distribution of bladder tissue data images. After data preparation, the existing CNN like DenseNet121, ResNet152, VGG19Net, InceptionV3Net and proposed 17-DCNN model is chosen by fitting the Endoscopic bladder tissue images. The accuracy of the proposed 17-DCNN was 99.32% when compared to other CNN models after implementation.