AnnonTip-AI

With this application, a user can report an incident/crime annonymously by leveraging AI. A reporter can add image or audio, and we will give you insights or transcription of that image or audio.

Comment

Inspiration

Reporting crime has always been easy with anonymous tipping. But it lacks in a lot of features like getting audio/image/video considerations. Mostly the existing system relies on descriptions from users. This project idea originated from the fact that images/audio can also give us a lot of insights.

What it does

  • Use images to get insights of a person like a face detection/cropping of face, age gender, and facial features like sunglasses, beard etc.
  • Use image to get moderation insights like Fire, Disturbing visuals, Weapon Violence, Gun, Drug abuse and much more.
  • Add Reports by adding description, zipcode etc
  • Reports are open to view for all

How we built it

  • Streamlit for frontend
  • Python API for creating the UI
  • Amazon EC2 to host this

Challenges we ran into

  • Getting the necessary insights from images were difficult at first. AWS offers a lot of image detection and labeling services. Finding something that would fit the project took considerable thought.
  • Assembly AI transcription
  • Data facing and usability in scale

Accomplishments that we're proud of

  • Getting it done
  • Making it easy and simple

What we learned

  • Building apps from scratch
  • Design a system in a product oriented way
  • Ideation and expanding
  • Limitations of the system

What's next for AnnonTip-AI

  • Use Maps for insights of location as per zip codes
  • Scale for considerations like video and multiple images
  • NLP on audio transcript and summarization of audio

*please find the loom link for product video demo

Built With

Share this project:

Updates