The problem Accist solves

95% of the lives lost during car accidents is due to the lag of time during reporting and getting medical help. we are trying to minimise this time by directly detecting the accidents from the CCTV cameras installed at every corner of the road. As soon as the live feed detects the accident it will give alert on the iOS app with location and image of the accident(Just to provide Human Supervision). You will see every accident there in the app and can even click on resolve button to delete it from the list if you find that the problem has been solved. Kindly see the Google Drive link for live demo.

Challenges we ran into

What we did is combined Yolo with a Classifier that can detect a crashed car from an image. And that combining part was just like hell. The second most difficult task was to integrate our model with iOS app. But at last we succeded in all the tasks and made a perfectly working model.