Digital surveillance and artificial intelligence are very closely connected. Today, scope of digital surveillance has gone far beyond the usage of security and safety. Data generated by intelligent IP security cameras are being used in various fields to improve automation with artificial intelligence. We had come across a few such applications way back in 2005. One of them was text insertion into video output. We superimposed receipt data of point of sale system on the video output of the camera. This was done to ensure that items delivered are the same as what the customer paid for.
Then I saw a brilliant use of IP cameras at the Wellington airport. Some 3-4 years ago, you needed to flash your RF ID card at the entry to come in the car park and at the exit to go out. If you remained at the airport for less than 10 minutes, you are not charged. But now, you don’t need to flash your card at the exit because cameras at the entry and exits have capability to capture number plates of the vehicles and convert the captured image into alpha numeric numbers.
IP Camera in Advertising for Split Testing
I came across a new usage of IP cameras last week, in reference to digital surveillance and artificial intelligence. We received an order from an billboard advertising company and they explained to us their requirements. We suggested 5MP Outdoor, Low light Geovision cameras. They needed 2 cameras. One looking at the billboard and other looking at the faces of the people watching the billboard. It is possible to alter the Ads on this LED based billboard remotely.
I don’t have much idea about various aspects of advertising but because of my involvement in eCommerce, I know about split testing of Ads. In split testing, you create campaigns with 2 or 3 variations of your Ad and run them simultaneously. After a predefined time (a week or a month) you choose the variation of the Ad which performs the best. These people were trying to achieve something like that. I think, they will run variations of the Ads, and by looking at people’s face expressions (their reactions), they will decide which Ad is generating the most positive results.
Above in the picture gallery, you can see the camera just below the billboard looking at the people watching the Ad and the camera at the far end is looking at the billboard.
Video below give a very clear idea about the the whole setup.
I explained above what ever I could understand from the perspective of the advertising world but the owner of the Bekon Media later told me that they are going to use additional software to minutely observe the face expressions.
Digital surveillance and artificial intelligence have a long way to go. Their incorporation in various robotics based projects will provide a new dimension to artificial intelligence.