Inspiration

We were inspired by the issues of public safety that are prevalent in Brampton and many communities alike that have rising trends of crime rates, and were determined to find an innovative solution using new technologies, such as a voice recognition software. Statistics of the Rising Shooting Rates in Brampton

What it does

Our program uses a voice recognition API to analyze an audio recording to detect certain key words. When they are detected, an operator is notified to check the full transcript, and a text is sent to civilians to make them aware of the possible threat.

How we built it

Used several APIs and front end libraries including Twilio for connecting with the public via text, pygame to wrap, speech_recognition, PocketSphinx. Mimicking a CCTV or writable feed as our microphone integration, the industry and common cost effective implementation at this time, we evaluate the queue of hot words against words recognized, also with specific phrase combination integrated within our algorithms to detect possible threats.

Challenges we ran into

A challenge that we ran into was choosing the correct energy threshold for detecting words, as it would sometimes detect words incorrectly in a loud environment. Another challenge that we encountered was creating UI for the program, as it was difficult to format all the buttons and text to represent the correct functions in the program.

Accomplishments that we're proud of

We are proud of the integration of the SMS notification system into our program to alert citizens of a possible threat in their area, and we believe that this would be a useful addition to enhance the functionality of our program. We are also proud of our website that we created, which helps to effectively convey our mission statement to the public, provide an overview of the program, and the positive effect that it can have on the community.

What we learned

We learned how to integrate a voice recognition API into our program, and learned the specifications for sound analysis (Hertz, Chunks, Sensitivity, and Sample Rate). In addition to that, we learned how to use the Twilio API, and integrate it into our program in order to send text message to civilians to alert them about a possible threat in the area.

What's next for Safe and Sound

We think that with some refinements and further development, Safe and Sound can be implemented into the surveillance systems of communities all around the world, to prevent crime rates from rising, and keeping the peace in our communities.

Visit our website

https://rodolfofunes.github.io/safe-sound/

Built With

Share this project:
×

Updates

Kevin Wang posted an update

The idea of automatic surveillance as opposed to an operator is almost the same in terms of effectiveness in protecting citizens, except you're able to run multiple instances, at almost no cost (eliminate a live workng, present listener per audio feed), except if you'd like a reviewer reviewing when a flag is said. But even then, they are able to work remotely, on call, and are useful only if algorithm and detection error must be eliminated to instigate appropriate action (eg. alerting first responders as opposed to just collecting data, etc.)

Log in or sign up for Devpost to join the conversation.