Friday 17 February 2012

CCTV Images: Looks like me, sounds like me but it is not!

Well, as usual, I’m not going to comment on any controversial CCTV videos or images…whether that person is really the person.

Nonetheless, as a researcher, I would like to share on an interesting article by Wilkinson et al. [1]. The title is “Are facial image analysis experts any better than the general public at identifying individuals from CCTV images?”

Interesting topic isn’t it?

In reality, whenever there is a CCTV case, the general public would be the one who will speculate; “I know this person…this guy…this lady.”

To make it worse, with the advent of Internet and YouTube…CCTV images and videos can go viral…in split second. And more and more people will speculate.

And, the reporters may make assumptions and decide to print the thoughts of the general public (I’m not trying to offend anybody here…it is purely based on experience and observation).

This can be bad. Is this good for the judge or jury?

Certainly not when you read the outcomes of Wilkinson paper.

According to the authors, facial image analysis experts are better because they are more consistent. Their identification rates are double and errors are half than the general public.

Experts are trained and their assessment is based on science. They will make analysis according to proven method and experience whereas the general public is on their naked eye. If the image is of high quality, the general public may get it right if not…their opinion may not be considered in the court.

Thus, the expert witness or attorney in the court of law could cite this empirical study (using ANOVA with good sampling).

Better isn’t it? Forensically sound! Scientific assessment!

[1] C.Wilkinson and Raymond Evans. “Are facial image analysis experts any better than the general public at identifying individuals from CCTV images?” Journal of the Forensic Science Society, vol. 49, pp. 191-196, 2008.