Contactless Covid Test & Personalized Diet Suggestions - AI

Web App integrated with AI systems to support Contactless COVID19 Test, Patient Prioritization, Personalized Diet Suggestions and COVID Chatbot.

Comment

Inspiration

COVIID-19, the most exasperating topic spoken all across the Globe. We are all aware of this virus and its consequences. People get afraid due to the symptoms of COVID-19 and crowd at the hospitals which is currently very difficult for the frontline workers to manage and take tests for all. It is very important to know who actually needs to be tested. Hence people who are at a higher risk are only needed to be targeted.

Also it has been found that the best medicine for COVID-19 is improving our immunity. So, a personalized diet suggestions for encountering COVID would improve each and every one to win this battle against COVID.

Lot of fake news about COVID are spreading on social Media where people get confused on what to do. So it would be better if they get answers for their questions on what to do and what not to do.

How this application works

A simple Web App integrated with AI systems is implemented which helps the Doctor to diagnose the patients without visiting them. The Web App contains the following:

  • Contactless Test
  • Personalized Diet Suggestions for COVID-19
  • Chatbot

Contactless Test

A risk score is generated by the AI system which ranges from 1 - 100. It is this score that helps to prioritize the patients. Higher the risk score, higher is the importance given to that patient.

The Contactless test consists of the following:

Dry Cough Severity

The user is made to cough and its sound is recorded and sent to the Database. The Convolutional Neural Networks reads the input frequencies using the spectrogram images and predicts the dry cough rate based on how severe the dry cough is. If its a normal cough or a wet cough, the dry cough rate would be less. This test contributes to the risk sore.

Heart BPM | Respiratory Test

The user is made to take deep breath in and out for 15 - 20 seconds which is recorded by the camera. The Convolutional Neural Networks read the input image sequence and generates the respiratory rate and bpm. Based on these values, the AI system keeps a track to the risk score.

Queries

Well framed queries are questioned where the user need to answer. Eg: "What is your Current Location?", Do you have any travel history in the past 30 days?, Are you able to smell odorous items? etc..,

The AI system itself finds how severe COVID is affected in the user's current location and if its high, then accordingly it gets tracked to the risk score. In the similar way all other questions are also tracked by the AI system to the risk score. Finally the AI System generates a risk score for the patient. Now Doctors can login the Web App at any time and can diagnose all the patients based on these tests. Finally the Doctor writes the prescription/remarks for the patient.This is sent to the patient's mail. As a prioritization is made, highly risky patients are diagnosed first. The user can follow the guidelines as prescribed by the Doctor in the Prescription.

AI Personalized Diet Suggestions for COVID-19

The users need to answer the questions which helps to calculate the magnesium, zinc,fat,water,sugar and vitamin content in their body. This is sent to the database. Random Forest Algorithm is used to determine what kind of diet is needed for that person in order to improve the immune system.

Chatbot

Its an RNN based chatbot. People can ask any kind of queries regarding COVID. All the data's and facts about COVID were fetched from official websites like WHO. So people can ask like "How to go to a grocery shop?", "Should I use masks at home?", "Are kids affected equally ?" etc...,

How this application was developed

  • The COVID Test, Chatbot, Diet Suggestions were created using the tensorflow and keras frameworks. (Python)
  • The UI was entirely made using HTML, CSS, Javascript.
  • A MYSQL Database was made locally using phpMyAdmin.
  • The AI system, UI and MYSQL was integrated using flask.

How to use this application

The packages mentioned in the Github link of the project has to be installed. Once the Database is created as listed in Github, run "python app.py" in the Command Prompt which opens the Web App.

Difficulties & Challenges we faced

Detection of the ROIs(Region Of Interest) in our face to calculate the BPM and Respiratory Rate. Converting the sound data into spectrogram images to train the CNN.

Accomplishments that we're proud of

We have used AI technology that can simplify the jobs by concentrating more on the highly risky people rather than taking tests at hospitals for everyone. This prioritization focuses on the early identification and protection of vulnerable patients and health care workers. Apart from that, a diet recommendation and Chatbot is made to effectively make an ALL-IN-ONE helpful website for the people.

What we learned

  • An in depth analysis was made to understand and evaluate the Nutrient levels in the patient and recommend them a suitable diet.
  • Working in Face Detection and finding the ROIs.

Go-to-Market Evaluation

The Database has to be pushed into the server. Once the Web App is hosted, it can be assessed by all.

What's next for Contactless Covid Test & Personalized Diet Suggestions - AI

Working to create a pulse oximeter which uses the image of our finger to detect our oxygen saturation levels in our blood. This is very useful for COVID test as it has found that most of the COVID patients have SO2 less than 90%.

Built With

Share this project:

Updates