AI, Video Analytics and Crime Prevention
Currently most city's police are installing more than a lakh cameras in public places. Let us see how this massive data can be used for crime prevention. Crime investigation maybe more or less straight forward and need little discussion. Let us concentrate only on crime prevention.
The mere presence of camera on the streets and public places will instill a fear in the minds of criminals. So they have to put some extra work or resources to escape the eyes which never blink or sleep.
Artificial intelligence based facial recognition have started behaving better than human beings. This field is not new. They are around 60 years old. Scientists were doing facial recognition using massive computers of a few MHz operating frequency and around 20K memory. The cost of those computers are not known, but none of the universities in the US can afford it in those days. IBM is the only supplier who can make it.
Currently AI based facial recognition tools can even identify a face more accurately even in a poorly lit conditions. Generally AI takes various measurements of a face namely, the distance between eyes, eyes and nose, size of temple, shape of the face etc. At the least it can give a list of 10 people who faces matches the test image.
AI is good in alerting unusual activities. Activities which does not fit a pattern. For example, most of us follow a fixed pattern of movement. We leave home at fixed time and mostly come back at more or less the same time. Any change in this pattern will indicate there is something wrong. It may not be a criminal activity. Some one may be ill at his home and his movement pattern might have changed. This can be identified if he is visiting a hospital. If he is not? Then it may need a detailed analysis.
If a group of unrelated people, either by profession or relation, started meeting at various places or at a same place, can be detected by AI. This is impossible by human beings. If such things are detected, we can start further analysis of those 10 peoples. There are software tools available which will analysis social media inputs. So one can check whether these people belong to same ideology.
If this is positive, then one can check their call history and even bank accounts of these people for any unusual deposit or withdrawal of money. One an analysis any unusual movement of material in the vicinity of their place of work or residences.
One can monitor the movement of materials too. Delivery of large parcel to residential area can be suspicious, provided the material is not from nearby consumer durable shop or furniture shop. It may not be criminal activity, but it could be unaccounted business activity, which is again is not legal.
It becomes easy to keep track of known faulters. One can keep track of people on parole. One can keep track of foreign national whose activities may be dubious, alerted by Interpol or similar agencies.
This facility can monitor the whereabouts of police force and can be quickly deployed to handle any emergency situation. Generally before any terrorist attack on a particular place and time, the terrorist usually will visit the place for better planning and execution. A close monitoring of all the people and their reassembly in any particular location, their past history of crime will reveal a lot of details and can be used to prevent any undue event.So crime prevention is becoming more of office activity than a street activity. People can prevent and investigate crime by sitting at the cozy office.
About the Solution Architect - Kadal has filed 4 patents in US and out of which 3 are awarded. Currently he is working in the field of IoT. He has current developed IoT products which have been in implementation.