Our inspiration comes from seeing the Covid-19 story unfold and wanting to help fight the fight. We saw firefighters, ambulances, and police cars and saw just how fast they are able to reach the scene. Using this technology, we can shorten this response time even more. Just in a few seconds, we can make sure they are able to provide better and safer ways to help.
What it does
The software makes it easier to recognize the patient. We can save time in helping first responders identify and gain access to a person’s basic medical information by encouraging users to keep a record of all of their information and log any symptoms they experience.
Information that can be recorded:
- User's phone number
- Face recognition
- Health and Medical records
- Location history
Health and medical records that can be logged:
- Blood type
- Activity record
- Heart rhythm
- Body temperature
- Emergency contact information: - Close persons - Health physicians - Insurance company (if applicable)
- Any other notes the user would like to add
This can be really important for people who don't have a medical record, or who live in areas and countries without medical records or medical insurance. Smartwatches can be used to detect and have records of heart rate, temperature, activity, and location history throughout a time period, and are saved in the user's record in the software. Users can log symptoms to track daily changes, - self-reported plus automatically reported - collected from devices such as phones and watches. Server, cloud, and all information are encrypted using military standard encryption, AES-256 bit key.
location-based tracking of corona cases reported from users and authorities. Include a geofenced area of possible contact. stay informed about family and friend's health status. chat and share features supported. users can draw and mark dangerous/safe zones for family and themselves and receive notifications if one of them enters or exits the zone. It will be available on Google Home, Alexa, and other voice assistant devices. Drones are another device that will be supported in order to increase the Covid-19 identification speed.
How we built it
Amazon SageMaker is used to train deep learning models on Amazon EC2 P3 instances. API Gateway connects to the AWS Lambda function, and Lambda Function connects to s3 and SageMaker. The application is built using Swift and Android. The application opens the camera, which can see the face of the user, and will take a video. The application will take 25 frames and upload it to AWS through API Gateway. Then the user will be detected and their information will be sent back to the application.
Challenges we ran into
One of the biggest challenges we had to overcome is reaching a face recognition accuracy percentage that is high.
Accomplishments that we are proud of
We're able to get about 90-95 percent of faces recognized accurately.
What we learned
We learned to use the machine learning along with several technologies such as geofencing, google maps, and many of Amazon cloud web services.
What's next for CoronaFront
We hope to be able to make the system available globally and continue working on face recognition accuracy.
amazon-web-services, android, firebase, google-maps, google-web-speech-api, laravel, machine-learning, nginx, oracle, s3, sagemaker, swift, ubuntu