CHANDIGARH: When does a mob turn violent, especially in areas hit by insurgency? Is it possible to visually detect whether a person is under the influence of drugs or alcohol?It will soon be possible for the police and other law enforcement agencies to accurately predict whether a mob could indulge in violence using a camera app that runs on special software created by IIT Ropar’s department of computer science and engineering. The software for mob mood analysis is at commercial stage, meaning it is now being developed for wider public use. However, the software to detect drug use by people is still at an early laboratory stage and will take time to hit the markets.The visual mood analysis app has demonstrated 80% accuracy in laboratory trials, which is considered to be reasonably good.The software can scan faces in a mob and check place and the scene around to understand the mood of a crowd. “It’s a camera-based real-time application. We started working on it in 2014 and it is funded by the IIT Ropar . Presently, it is a desktop-based app and we are working to make it android based,” said Abhinav Dhall, assistant professor in IIT Ropar’s department of computer science and engineering.How can this system of surveillance help government agencies? Cameras can be attached to traffic lights in areas hit by insurgency and the data collected can be monitored for real-time feedback to a central repository. “Any collective negative emotions can be sensed for timely action. It needs just surveillance cameras and a desktop. An MNC has come forward for making it available commercially,” said Abhinav.The application can also sense mood of a crowd watching a match or a movie. “We can also use this for selecting candidates during an interview,” added Abhinav.The software is based on deep learning — an artificial intelligence function that imitates the workings of the human brain in processing data and creating patterns for use in making decisions.“Some years ago, MIT experts had been working on group-level emotion recognition artificial intelligence-based software. But there is no such developed product in the market till now. There have been some publications from the universities of China and the US (on the issue),” said Abhinav.Using the same technology is another AI-based software which can sense if a person is intoxicated or on drugs. “We wanted to address local issues and work towards some solution. So, this software can not only detect if a person is on drugs, but also at what stage of addiction is the person involved. As this software is non-invasive, it is not harsh on the addicts,” said Dr Sarit Das, director, IIT Ropar. “We find it very relevant for Punjab where drug addiction has been a problem.”The software analyses head, eye and upper body gestures of drug addicts and non-drug addicts and then uses an algorithm to pick up nuances and differentiate if a person is intoxicated or not, the experts said.The AI software for drug addiction can be used by law enforcement agencies and parents as well. “There is a software developed by a German group which is speech-based. Ours is mobile-camera based,” said Abhinav, adding that “if we get sufficient funding, we can offer these applications free of cost.”The software can be connected wirelessly to any surveillance camera. As it is a camera-based application, images captured will be analysed by the AI programmed to understand mood and emotions in a crowd. The data after interpretation is send in real time to a central repository and the application highlights the mood of the crowd.Similarly, the body gestures can be interpreted in case of intoxication analyses. The drug addict or a person intoxicated will be asked to look at the mobile camera and speak for 20 seconds. The software uses body gestures of the person who is speaking to analyse and detect level of drug addiction or intoxication. To the end mobile or desktop user, the software will give the state of the emotion of aggressive, calm, sad or in state of intoxication.

Stay updated with all the Chandigarh Latest News headlines here. For more exclusive & live news updates from all around India, stay connected with NYOOOZ.