Skip to content

project.md

Latest commit

 

History

History
19 lines (14 loc) · 1.13 KB

project.md

File metadata and controls

19 lines (14 loc) · 1.13 KB

Problem statement

There is CCTV almost everywhere but their primary role, is to store videos for future reference(evidence) just in case something happen. Sometimes, those who commit the crimes cover their identities, so they can't be traced down while victims has already suffered the damage. In Korea, crimes rate is low but that being said some neighbourhood like Itaewon or hongdae, crimes like sexual harassement, assult are frequently reported. Globally, crimes such as shooting,rape,sexual harassement, assult, and so on. Our goal is, as "prevention is better than cure" instead of using CCTV for evidence only,then how about using them to help in encountering such crimes when they are happening? With help of Artificial intelligence.This goal can be achieved.

Problem solving strategy.

  1. Train model to recognize and classify crimes, deploy model.
  2. After recognizing crimes and it's severity, alert nearby police station or other security agent.
  3. Police will be able to respond to the scene instantely.

Models to consider

  1. Video classification model trained from images?
  2. Video classification model trained from portion of videos?