Friday, April 19, 2013

Image Analysis Domains

With every US news outlet scrambling for a unique angle on the bombing in Boston, and with the 24 hour cycle, news producers have been trolling the internet for "experts" to keep the story alive. I've seen a few of the reports, and I'm quite frankly embarrassed at some of the reporters' questions.

The main point of confusion, on the part of the media, is the word "analysis." In the context of this discussion, the definition of forensic video/image analysis can be found in LEVA's glossary:

Forensic Video Analysis
(FV) - (definition developed and accepted by practicing forensic scientists/examiners) Forensic Video Analysis is the scientific examination, comparison, and/or evaluation of video in legal matters. This is also called Forensic Video Analysis. (ASCLD-LAB - category definition for accreditation purposes only) A subdiscipline of Digital Multimedia Evidence which involves the examination, analysis, and/or evaluation of video.

The key word in the entire block is "scientific." In order for the analysis to be "scientific," it must follow the scientific method. The large themes of the scientific method involve testing your hypothesis against logical alternatives, asking for and receiving peer review, and the academic underpinnings of your theory. Is your hypothesis testable? Does existing literature support your test? Can your peers reproduce your results in an independent setting?

In terms of the domains of analysis, there are four agreed upon domains for forensic video/image analysis:

  1. Photogrammetry
  2. Comparison
  3. Content analysis
  4. Authentication
One common mistake made across the news reports I watched is that clarification/enhancement is analysis. To be sure, clarification/enhancement help observation and make it easier to form initial conclusions. To be part of a scientific method, they need to be reliable and repeatable. But they are not, of themselves, analysis. 

Of the many "experts" who accepted the opportunity to speak to a national audience, understand that the opposition research team for every case you will be involved with in the future will find your statements on the news sites and on YouTube. You will be asked about them. That you were taken out of context or selectively edited won't matter much. I know this well as I've written on the subject for years, presented at countless conferences, have had a few appearances on TV, and have a couple of books out there. 

I know that the media doesn't care about me or my reputation - they care about ratings and keeping an audience. By the next day, they're on to the next big story. But, your comments now live forever on the internet.



No comments: