Inspiration
The Coronavirus pandemic has closed thousands of schools around the world, in fact 1.4 billion students are now at home. Many parents now have to juggle providing full-time childcare while working or being full-time students themselves. SuperAlvise aims to alleviate this burden, by providing parents the peace of mind to focus on their studies, while knowing that their children are safe.
What it does
SuperAIvise uses a Machine Learning algorithm to track the child's movement using a webcam. This is done securely to avoid privacy risks. When the child leaves the area, an SMS is sent to the parent so they can check on the child. If the child wants to alert their parent they can block the webcam with their hand.
How I built it
I built the web application using Django and HTML and CSS for the front-end. To identify and track the children I used the OpenCV, ImageAI packages alongside the YOLO image database. I also implemented Twilio to send messages to the parent
Challenges I ran into
I found it very difficult to get TensorFlow and ImageAI to work together, I had to try multiple different versions of each package and Python before I found a winning combination. The integration of the webcam feed into a Django web page was also very difficult and not very well documented.
Accomplishments that I'm proud of
I learned how to use threading in python, this enabled me to show the webcam feed to the user and track the location of the child at the same time. I was also pleased with the look of the website, I think it looks professional. I was not particularly experienced in front-end development but I really enjoyed developing my HTML and CSS skills
What I learned
How to use object detection software, this will be very useful in future ML projects
What's next for SuperAIvise
I want to add optional boundaries for parents to add, so they can choose which areas of the room they want their kids in. I would also like to add educational resources to the application, for example, classes or programs that children could complete within SuperAlvise.
Log in or sign up for Devpost to join the conversation.