Photographic image comparison is the process of comparing object(s) or persons(s) when at least one of the items in question is captured in imagery, and making an assessment of the correspondence between features of the captured imagery for rendering an opinion regarding identification or elimination (as opposed to a demonstrative exhibit). Examples of photographic comparisons include, but are not limited to:
- A facial comparison between an unknown subject depicted in a surveillance image with an identified suspect.
- The comparison of objects such as vehicles depicted in surveillance images with those recovered in an investigation.
Any methodology applied to photographic comparison should incorporate an analysis of the imagery, a comparison of individual features, an evaluation of the significance of the comparison and a verification of the comparison.
-SWGDE Guidelines for Forensic Image Analysis PG 6
The Scientific Methodology
The photographic comparison investigation science has been established by the scientific community through SWGDE (Scientific Working Group on Digital Evidence).
An expert like myself can perform a forensic image comparison investigation if the methods applied approved by the scientific community. The expert must also be trained by an accredited agency in the scientific method. I have received training in photographic image comparison through LEVA (The Law Enforcement and Emergency Services Video Association).
The photographic image comparison investigation is not bound by images extracted from surveillance equipment, but images of sufficient quality. These images are created by many types of digital camera systems. These images must be of sufficient quality to analyze and extract characteristics for the photo forensics analysis.
It is considered that today, the most prolific source of evidence for police is from video images. Video is not an accurate medium for identification on its own. This is because digital compression removes and adds information that wasn’t originally included. An Image Forensic Expert MUST be able to interpret the data to provide an accurate digital image comparison investigation.
Because of image unreliability of images, an Expert MUST also be able to qualify the events based on quality. This is done to ensure that the objects or people presented in them are accurate. This includes sufficient detail and pixel information. In some cases when the objects or people displayed in the images and videos aren’t clear, we will perform a forensic video enhancement investigation. This may become necessary in order to see the details of the images more clearly. The video enhancement process will also reinforce the accuracy of the comparison investigation.
Exemplar VS Evidence
Evidence: Images or videos that contain objects or people that are considered UNKNOWN
Exemplar: Images or videos that contain objects or people that are considered KNOWN
Often times when an exemplar image or video is unavailable, we have to make other arrangements. An acceptable practice is to gain access or obtain similar equipment that was used to create the evidence. This equipment must be programmable to similar parameters that were used at the time that the evidence was recorded. This ensures the highest accuracy when comparing the images or videos forensically. If exhibits for courtroom use was generated using different equipment, it isn’t necessary to explain why. For that reason, if the details and events captured are sufficient in quality for comparison, this is acceptable.
These images must be compared using acceptable best practices approved by the scientific community. These best practices include the acceptable criteria for comparison. The criteria are divided into two categories, class characteristics and unique characteristics.
Class characteristics: A class characteristic is an identifiable feature that assists in narrowing the statistical probability that a questioned object or person belongs to the same group as a known object or person that shares the same feature. Consistent class characteristics between questioned objects and individuals can only suggest similarities and cannot, by themselves, be used to infer a positive identification.
Unique Characteristics: A unique characteristic is an identifiable feature found on the questioned object or person that is found on no other object or person of similar class. A unique characteristic can be used to significantly narrow the probability that a known object or person is in fact the same questioned object or person. A thorough examination of class characteristics and unique characteristics contrasted and compared between questioned and known objects or persons may allow an experienced and properly qualified analyst to form an opinion. If it is safe to do so, the objects are one in the same, eliminating all others of the same class.
The comparison criteria is defined by the standards presented by LEVA (Law Enforcement & Emergency Services Video Association).
Expert Witness Testimony
It is our responsibility as an expert witness to present observations in a simple unbiased nature to the trier of fact. When we write our reports, we ensure that the foundation our opinions formulate from is scientific. This gives the testimony and report integrity, therefore making it “bulletproof” if you will. In other words, we do not generate opinions about the likelihood of objects or people based on a hunch.
An example of a hunch could be a detail open to multiple interpretations or from digital compression artifacts. It is also acceptable to work with and consult other qualified experts. For example, if identification of a specific type of vehicle is needed, an automotive expert can be retained.
These principles strengthen an investigation to involve another expert that has additional experience and qualifications to make specific determinations about our findings once we have completed the comparison investigation. This strengthens our reports as well as our testimony.
Click HERE to email your questions or
Call 800-647-4281 in the USA or +01-248-853-4091 Internationally.