Enhanced Electronic Diagnosis System by CCNNViT Model for Kidney Cyst Detection

A major health issue for kidney failure is becoming worse due to changes in diet, obesity, underlying medical conditions, mineral deficiencies. The kidneys calculus, also known as kidney cysts, is a hard accumulation of nutrients from the stool that develops in the urinary tract. One method of imaging that medical professionals use to diagnose kidney cysts is computerized tomography (CT). Kidney cysts are sometimes difficult to identify with the unaided eye owing to the poor quality that produces these images, which could result in false alarms. In order to help clinicians with diagnosis, an advanced deep learning technique-based electronic diagnosis system has been built throughout this work. Kidney cysts can be identified by CNN based deep learning algorithms. until now, it has issues with the conventional implementation and the efficiency of the convolution operations performed in the convolution layers. The suggested deep learning architecture uses a hybrid approach as a combination of CNN and pooling vision transformer (CCNNViT) technique that is incorporated in the deep learning architecture to decrease characteristic map redundant employment without the convolution overlap. We propose a method that extracts stated and abstract characteristics from the input pictures, thereby improving the network's efficiency. The suggested architecture is developed using the kidney cysts CT scans that are made accessible to anyone on Kaggle. Using CT scans, our automated simulation identified kidney cysts with a 98.9% accuracy rate. Our method outperforms some of the most recent and modern techniques for detecting kidney cysts of any form, even the tiniest previous ones. The present research proposes an innovative method for detecting cysts in the kidneys that uses Pooling-based Vision Transformers (PiT) and excludes the uterus from the evaluation.