IIT Madras students takes preventive measure in Riots and Stone Pelting using AI

In India, Protest is not just a word, more than emotion, its part of daily life. Some people protest for months in silence and make their way in amending the causes of the problems or even shutting down factories.

But, Politicians, Lawmaker, Lawbreakers don't like such valid questions in many cases. They have several precautional steps towards avoiding these protest like having a dispute between two groups.

At some stages, Protest would eventually become riot causing death without any probable reason and hunting people down have been a fashion altogether in many places.



Accusing the people in a major riot is almost impossible but IIT, Madras students came up with a solution of using Artificial Intelligence to find the people behind the scenes.


The Usage of CCTV cameras obviously found increasing over the last 3 years since Indian Police largely recommending to make it mandatory at maximum places to secure evidence.

IIT students use predictive modeling based on an action recognition algorithm in the video file captured out of CCTV. This project has been dedicated to the Indian Army, especially to the Kashmir, where stone pelting is a regular dilemma.

 "Intelligence - Finding Right Facts"

The abnormal action recognition using AI can also have a layer of YOLO project which predicts with moving images and also works in real time.

YOLO(You Only Look Once) - A Real-Time Object Detection is a git-hub package -Link here

This library has a trained dataset which would have been replaced by Abnormal and normal action sequences to make a predictive model which on the feedback loop gets updated with Machine Learning Algorithm in improving accuracy on its own.

Live images + Crowd Density map + Abnormal Action - predict the stages of breaking out a riot and preventive measures will be done.

This article helps in knowing about both YOLO project and real-time application of using Machine Learning to solve a problem in complex situations.

Hope... it's useful!!!





Post a Comment

0 Comments