We as a team shared the same interest in knowing more about Machine Learning and its applications. upon looking at the challenges available, we were immediately drawn to the innovation factory and their challenges, and thought of potential projects revolving around that category. We started brainstorming, and went through over a dozen design ideas as to how to implement a solution related to smart cities. By looking at the different information received from the camera data, we landed on the idea of requiring the raw footage itself and using it to look for what we would call a distress signal, in case anyone felt unsafe in their current area.

What it does

We have set up a signal that if done in front of the camera, a machine learning algorithm would be able to detect the signal and notify authorities that maybe they should check out this location, for the possibility of catching a potentially suspicious suspect or even being present to keep civilians safe.

How we built it

First, we collected data off the innovation factory API, and inspected the code carefully to get to know what each part does. After putting pieces together, we were able to extract a video footage of the nearest camera to us. A member of our team ventured off in search of the camera itself to collect different kinds of poses to later be used in training our machine learning module. Eventually, due to compiling issues, we had to scrap the training algorithm we made and went for a similarly pre-trained algorithm to accomplish the basics of our project.

Challenges we ran into

Using the Innovation Factory API, the fact that the cameras are located very far away, the machine learning algorithms unfortunately being an older version and would not compile with our code, and finally the frame rate on the playback of the footage when running the algorithm through it.

Accomplishments that we are proud of

Ari: Being able to go above and beyond what I learned in school to create a cool project

Donya: Getting to know the basics of how machine learning works

Alok: How to deal with unexpected challenges and look at it as a positive change

Sudhanshu: The interesting scenario of posing in front of a camera while being directed by people recording me from a mile away.

What I learned

Machine learning basics, Postman, working on different ways to maximize playback time on the footage, and many more major and/or minor things we were able to accomplish this hackathon all with either none or incomplete information.

What's next for Smart City SOS

hopefully working with innovation factory to grow our project as well as inspiring individuals with similar passion or desire to create a change.

Share this project: