Outrageous Claims Hurt the Industry

I just watched a video clip of an interview and demonstration of VideoIQ's abilities. While the first 3 minutes are standard analytic marketing, the ending of this clip is outrageous. These claims are, at best, wildly misleading, and certainly damaging to our customers and the entire industry.

[Update 2010: VideoIQ has removed the video and discontinued this marketing effort / sample case]

VideoIQ claims that they will be able to handle:

• Your child is lost in a crowd. It uses your image to find a shot of your child.
• Then it takes that shot of your child to find your child live.
• A subway bomber. Find a picture of a bomber, click the bomber and find pictures of anyone who looks like the bomber in the London transit system.

#1: You have to recognize how video cameras in malls and subways are positioned. Routinely, they are 8 meters high, ceiling mounted and covering a radius of 30 meters. 99% of them are analog. The image you get of any person is at a very poor angle, and the quality is usually quite poor. Yes, megapixel cameras are coming but this will not eliminate the angle issues. Angles are bad because shots of the backs of people head aren't very helpful when you are searching for 1 person out of 10,000.

#2: You have to realize how many people are in a mall or a subway. Even if your analytic is 99% accurate and you miraculously got perfect shots of all people, if you analyze 100,000 people, that's 1,000 false results. In a subway system over days, that's millions of images of people, meaning tens of thousands of false matches. Certainly it's better than scanning video tape but it's still a huge job even if you have perfectly positioned cameras - which you won't and will have a hard time justifying paying to install.

This ends up hurting all of us. Customers start with an expectation that cannot be met. They get disappointed, don't buy and tell their colleagues of the problems.