Safe Halo

Inspiration

We were inspired by the thousands affected by gun violence every year and one important statistic: the average school shooting lasts 11 minutes yet the average police response time is 13 minutes. So we thought of ways to combat this through Computer Vision recognition of weapons before an incidence occurs.

What it does

Uses a neural network to detect when a weapon like object enters the vicinity of a camera.

How I built it

We used YOLO (You only look once) , a convolutional neural network for image detection. We used pre-trained data for the weapon detection. Once a weapon was detected, we used twillio API to send a text message to police or first responders (in this case we sent it to one of our phone numbers)

Then, we created a website to showcase the information and release the code to the general public in the near future.

Challenges I ran into

Implementing Twillio API Setting weight percentages for the detection software

Accomplishments that I'm proud of

It actually works!

What I learned

Not everything works well on the first try. We failed multiple times with opencv but were able to overcome in the end.

What's next for Safe Halo

Possibly implementing it as a viable market-ready product for schools and businesses to purchase for their own safety.

Built With

• a-convolutional-neural-network-for-image-detection.-we-used-pre-trained-data-for-the-weapon-detection.-once-a-weapon-was-detected
• css3
• html5
• opencv
• python
• we-used-twillio-api-to-send-a-text-message-to-police-or-first-responders-(in-this-case-we-sent-it-to-one-of-our-phone-numbers)-then
• we-used-yolo-(you-only-look-once)
• yolo