Word of the Month – Biometrics

The National Academy of Science has advised forensic experts to deliver measurable and reproducible results when submit evidence and opinions in court. As a result, I’ve noticed that forensic science has a new FFB (favorite forensic buzzword) – BIOMETRICS.

 

So, what the heck is biometrics, anyway?

In the world of forensics it simply means, measurable characteristics used to identify an individual. As a document examiner, I could use handwriting biometrics to help identify an individual by measuring handwriting features such as the shapes and sizes of the writing. Biometrics can also include direction, pressure, location, punctuation and other individualizing traits left behind by a writer, depending on the samples provided and the requirements of the case.

 

Why measure, anyway?

That is the $50,000 question. Judges and juries would rather know in percentages how sure the expert  is about his opinion; attorneys want to understand just how helpful their evidence is; and, mathematicians love to create formulas (otherwise known as an “algorithms”) to provide answers to these queries. Most importantly, NAS wants to assure that forensic professionals base their findings on solid research. Unfortunately, everyone wants a quantitative answer for features that are presently unquantifiable.

 

In handwriting identification, there are a number of distinguished and respectable organizations, here and abroad, currently tackling the problem of “measuring the proof.” Most notably, in the US are two such organizations with two very different algorithmic approaches to identifying writers.[1],[2] Both are able to provide statistics and digital support for identifying a writer with a reasonable degree of accuracy using certain physical features of the writing. However, each has determined that most handwriting analyses are not complete without an examination by a properly trained handwriting expert. They agree that such things as handwriting subtleties and range of variation do not translate well to the digital world.

 

As I see it, computer science is serving my field well. Computers have proven that consistent application of a given set of parameters to handwriting measurements will provide consistent and reproducible results when identifying authors of handwritten text. More importantly, the conclusions of the computer scientists serve to further justify my professional existence by recognizing the importance of the expert as key in completing a handwriting analysis.

 

There is no denying that biometrics merits its FFB ranking among expert witnesses. It is important to me that the science of my field continue to grow and develop. My future as a handwriting expert depends on continuing validation by traditional and digital sciences. The one surviving truth that science cannot yet account for is that there is no substitute for the human element in the identification of handwriting. So for now, I continue to hone my observational skills and build my mental database of handwriting features; meanwhile, biometrics can be my FFB also.

 

 


[2]Center of Excellence for Document Analysis and Recognition (CEDAR) http://www.cedar.buffalo.edu/handwriting.html