TideWatch

Inspiration

We were motivated by the current UCSD reporting system. The current UCSD reporting system takes a while for an email to be sent to students' inboxes. Between the time a report is given and sent out, a student could have seen the incident and taken the necessary precautions to make sure they are safe. We wanted to create a real-time reporting system.

What it does

When creating a report, we ask the user where they are on campus and upload an image of what they see. This image goes through our EyePop.ai model which determines if there is fighting, law enforcement presence, someone wielding a weapon, etc. Based on the contents of the image, our website returns a risk score to determine the danger level of students around the area. The user is also allowed to provide a description of the situation and whether or not a suspect is visible in the photo.

How we built it

We built the backend with Firebase and created our model with EyePop.ai.

Challenges we ran into

The biggest challenge we ran into was trying to create a mobile application. Our laptops kept crashing and we ran into many errors so we decided to build a website. Also, finding databases to give create a sufficient model. We had to find many images of fires, law enforcement activity, people holding a weapons, pictures, of fights, and finally deploy the model.

Accomplishments that we're proud of

We are most proud of connecting our model to the backend because it was difficult to set up the API keys.

What we learned

We learned how to create and use our API keys and update our database using Firebase.

What's next for TideWatch

We want to feed the model more datasets so it could classify more potential harms and be more accurate when giving a risk score. Also, we would like to feed a database of pictures of the school so the model could recognize what buildings are related with the incident and where on campus a user is if they don't know where they are at.