COVID DETECTION FROM CHEST XRAY USING DEEP LEARNING AND IOT

Project Title

COVID DETECTION FROM CHEST XRAY USING DEEP LEARNING AND IOT

Project Area of Specialization Computer ScienceProject Summary

COVID-19 is currently a worldwide pandemic disease that needs to be  detected at early stages for proper diagnosis.Many medical experts have predicted that COVID-19 may prevail for years  before it can be completely eliminated . One of the major problems we face during the cure of Covid-19 is  its early screening and detection. Due to contactless spreading of the virus its screening is really really difficult. 

The health community needs effective means for identification of the virus in the shortest possible time. We invented a medical device consisting of composable sensors to monitor remotely and in real-time the health status of those who have symptoms of the coronavirus or those infected with it. The device comprises wearable medical sensors integrated using the Arduino hardware interfacing and a smartphone application. An IoT framework is deployed at the backend through which various devices can communicate in real-time. The data our system get from the sensor will then be send to our system and the tained system will predict the COVID-19 Result.

In addition, the proposed system will provide an android application for the general public to check their Covid status based on X-Ray. Users just need to upload an image and the proposed trained model will predict the result.

Our proposed system will be modelled specially for third world countries like Pakistan and will be easily  usable  . Our proposed System is based on Deep learning and IOT as well. It will help to screen out most people in a minimum time period.

Project Objectives

The main objectives of our proposed system (Covid detection from IoT sensors and  Chest X-Rays) are:

1. It will prevent the outbreak of the COVID-19 as we have sensors in public areas to check whether they are corona positive or not and we also make an android application for the general public who can’t afford these sensors so that they can easily check their corona positivity.

2. The proposed system will provide a more understandable and usable interface to the Users.

3. The proposed system will provide an efficient  way to stop the outbreak of COVID-19

4. It will solve the problem of buying expensive test kits of COVID-19.

5.  The proposed system will be time efficient, So we can screen out maximum people in less time.

Project Implementation Method

Our  proposed system is used for overcoming the problems of shorting of Testing kits, we proposed an initial framework based on IoT and Deep Learning. Our project has 2 modules one for single person use and the other for screening out in public gatherings.

The Module of Single person use is based on Deep Learning technology in which the user just needs to input his/her X-ray image and our train model will predict his/her Covid Result. Because everyone can not afford testing kits and Sensors and it will also save their time and their public exposure.

The other module is based on the Internet of things(IoT). In this we deploy different sensors i.e MAX30205 (for measuring human body temperature), SW-420 (for cough detection), MAX30100( for measuring pulse/heart-rate) and ESP8266(for wifi connectivity) and PT100.

A. Deployment of Sensors

The sensors will be deployed in the human for the collection of data about his/her temperature, cough detection and heart-rate.Project will be implemented using Internet Of Things and also image processing/recognition.Implementation method includes using several sensors for sensing temperature, cough and heart-rate.

B. Transfer of information

The microchip ESP8266 has the ability to connect multiple things in a Wi-Fi network. This is an essential requirement in our case as we are connecting multiple sensors in our solution

C. Analysis/Comparison

Our trained model will get data  in the form of heart-rate, cough detection and body temperature. Our System has 5 Convolutional layers, Input, Conv2D, Max Pooling, Flatten and Dense. Our proposed system will then generate results based on that data.

D. Recommendation to Users

Finally, after detecting the covid it will  send a message on the screen for authorities to Send the person in the COVID-19 centre and follow Sops.

Benefits of the Project

The COVID-19 virus has been around for almost two years now and the medical community, scientists, and researchers are trying their best to identify a cure for the disease. At the same time, people around the world are facing issues in determining the state of an individual as healthy or affected by the virus. The state-of-the-art solutions require visiting a hospital or a healthcare facility to perform COVID-19 diagnosis. Considering the numerous difficulties and the associated dangers in its diagnosis, it is preferable to be able to perform the disease detection using wearable devices.This project will provide several advantages including cost effective disease detection, accurate and fast results.

1. Low Risk as if the person is covid-19 positive and comes to a public gathering there is a high chance of virus outbreak.

2. Low rate of failure: as a human the rate of failure is high as compared to machines.

Technical Details of Final Deliverable

The proposed system will require an image capturing device, Arduino, and sensors (i-e temperature sensor, cough detection sensor, heart-rate monitor sensor). These sensors will sense certain data including temperature,cough variation and pulse rate.. The information gathered will be sent to trained model using ESP8266 associated with Arduino board. Then this data will be tested at the system. Finally it will generate the results either person is positive or not.    

Final Deliverable of the Project HW/SW integrated systemCore Industry MedicalOther Industries IT , Health Core Technology Internet of Things (IoT)Other Technologies Artificial Intelligence(AI), Wearables and ImplantablesSustainable Development Goals Good Health and Well-Being for PeopleRequired Resources

Item Name	Type	No. of Units	Per Unit Cost (in Rs)	Total (in Rs)
			Total in (Rs)	79500
SW-420	Equipment	1	8000	8000
max30205	Equipment	1	8000	8000
Arduino UNO	Equipment	1	3500	3500
max30100	Equipment	1	6500	6500
BeagleBone Black wireless	Equipment	1	25000	25000
FYP Documentation (Stationary, Printing, Binding)	Miscellaneous	1	9500	9500
PT 100	Equipment	1	19000	19000