Deep Learning Driven Kidney Stone Detection in Ultrasound Images Using Transformer and Graph-Based Models

Project Code :TCMAPY2135

Objective

The project aims to implement and compare various deep learning models, including Vision Transformers, MobileNet + GCN, and Inceptionnext + GCN, to detect kidney stones in ultrasound images. The performance of these models will be evaluated using different metrics such as accuracy, precision, recall, and F1-score to identify the most effective model for this task. The ultrasound dataset will be preprocessed and prepared for training, incorporating data augmentation techniques to improve model generalization. Additionally, a user-friendly web application will be developed using Flask, allowing healthcare professionals to upload ultrasound images and receive automatic classifications of kidney stones. Finally, the results from each model will be analyzed, and recommendations will be made on the most efficient deep learning model for kidney stone detection in terms of both accuracy and performance.

Abstract

This project focuses on the comparative analysis of deep learning models for detecting kidney stones in ultrasound images. The objective is to explore various deep learning techniques, including Vision Transformers, MobileNet + GCN, and Inceptionnext + GCN, to identify kidney stones from ultrasound images with high accuracy. The dataset used in this study is publicly available and contains ultrasound images labeled with "stone" and "no stone" categories. The models are trained and evaluated based on metrics such as accuracy, precision, recall, and F1-score. The Vision Transformers model leverages transformer-based architecture for better spatial understanding of image data, while MobileNet + GCN and Inceptionnext + GCN use convolutional networks integrated with graph convolution techniques to enhance feature extraction and spatial dependency understanding. This comparative study aims to find the most effective model for automated kidney stone detection. The results of this study can provide valuable insights for the development of automated systems that assist healthcare professionals in diagnosing kidney stones quickly and accurately.

Keywords: Kidney Stone Detection, Ultrasound Imaging, Vision Transformers, MobileNet, GCN, Inceptionnext, Deep Learning, Image Classification, Performance Evaluation, Automated Diagnosis.

NOTE: Without the concern of our team, please don't submit to the college. This Abstract varies based on student requirements.