Executive Summary : | Liver diseases are becoming a chronic health disease and mortality due to liver diseases is increasing rapidly. Most of the liver diseases do not show any significant symptoms at their early stage. Therefore, early detection and regular monitoring of liver diseases are required to prevent serious complications such as loss of liver function and liver failure. The main aim of this project is the early detection of liver diseases from the ultrasound images based on the internet of things (IoT) approach with the help of a developed Android application. The proposed computer-aided diagnosis (CAD) system designed using the deep learning models (DLMs) like residual learning network (RLN) and light weight-CNN (LW-CNN) helps to automatically classify the liver abnormalities present in the ultrasound images. Different liver ultrasound images like normal, fatty (steatosis), cyst, tumor, and cirrhosis of the liver are collected from the standard dataset and hospitals. The ultrasound transducer attached to the Android device will acquire the liver images. The obtained liver ultrasound images are pre-processed using RLN. Since the ultrasound images are usually acquired with the speckle noise. It is essential to remove the speckle noise from the ultrasound images. After the pre-processing step, the appropriate features are automatically extracted from the liver ultrasound images using the LW-CNN. The presented CAD system will be helpful in the automatic detection of liver abnormalities from the ultrasound images. The CAD system is designed in the form of an Android application which will be incorporated into a handheld Android device. The application helps us to monitor the condition of the liver and helps the patient to seek medical advice if the abnormality is detected, by suggesting the route to the nearby hospital. The presented CAD system will greatly reduce the burden of well-experienced hepatologists and gastroenterologists. This project is very much useful for poor people, mainly in rural and semi-urban areas. The CAD system would be beneficial in assisting the doctors in precise and rapid diagnosis. The designed Android application would serve as a supporting tool for diagnosing liver abnormalities, especially in the rural and semi-urban areas where the diagnosing facilities are limited. |