Featured Post

Welcome to the Forensic Multimedia Analysis blog (formerly the Forensic Photoshop blog). With the latest developments in the analysis of m...

Monday, November 19, 2007

Forensic Photoshop for everyone

Today's topic is not necessarily a step-by-step or a how-to. Today's topic is about communication. To make my point, I will be using images from one of our less than stellar career criminals.

Skippy (not his real name) decided to commit a series of robberies around town. Over the course of a few days, Skippy managed to rob six convenience stores; netting a total of about US$1,000.

Ever the fashion conscious individual, Skippy decided to wear his favourite set of clothes whilst on his crime spree. Luckily for us, Skippy's clothes are rather distinctive. The image settings allowed for on this blog are not the greatest for reproducing small DVR files, so I've helped us out by drawing a red box around the area of interest (use the Marquee Tool to select the area, then apply a Stroke - Edit>Stroke of about 4 pixels, red colour, location = outside).

There are some generalisations that we can make about Skippy's clothing. He is wearing what appears to be black high-top sneakers with white ankle length socks. His shorts come to about mid-calf length and there is a light/colour combination that is not consistent with a reflection and may indicate some sort of pattern or applied image. His tunic has a distinct colour pattern as well. Bear in mind that these are just generalisations - kind of like first impressions.

Skippy gets arrested

This is where it gets silly. Skippy is arrested later that week in the same clothes he wore when he allegedly committed the crimes (no longer allegedly, he is currently a guest of our Correctional System). A quick thinking staff member in our booking area thinks (wisely) to snap a few pics of Skippy's rather distinctive outfit. These get placed in the case file (great example of non-verbal communication).

Photographic / Video Comparison

Take a quick jump over to LEVA's site. You will find information on their next Comparison coursePhotographic / Video Comparison focuses on the science of comparing known objects, vehicles, clothing and humans with CCTV images of questioned objects, vehicles, clothing and humans. So here we have booking photos of known objects and DCCTV evidence from the crime scene.

Black high-top sneakers present in both scenes? Check.
Ankle high white socks present in both scenes? Check.
Blue calf length shorts present in both scenes? Check.
Possible distinctive pattern to shorts present in both scenes? Check.
Checked pattern tunic present in both scenes? Check.
(and the part not shown to protect Skippy: he's got a uniquely styled beard as well that shows up in both the booking photos and the DCCTV)

Now, there is much more to this as you will learn in the class. Suffice it to say, everyone in the process can benefit from a little forensic knowledge; the person who retrieved the DCCTV from the crime scene, the person who processed the DCCTV back in the police lab, the person who shot the booking photos, and the person who processed those booking photos for the case package. Everyone has a part to play; thus everyone can potentially benefit from proper training. So often, vital staff are left out of the training loop. When are are looking to spend those training dollars, we need to communicate this need effectively.

Use Photoshop's Arrange>Tile Vertically
(horizontally - depending on aspect)
to view the before and after shots on one screen.

As far as processing the images from the crime scenes was concerned, they were fairly straight forward. Same goes with the booking photos. The forensic part comes at trial; when you are going to argue your results. You have come to your conclusion; the clothing in the scene and the booking photos is the same (pattern, stitching, colours, consistent ... etc). There is what you know and there is what you can prove. Proving it requires not only your imaging talents but also the forensic skills of oration and debate.

When it comes to training, managers need to get adequate training for everyone involved in the process. If an employee is working with evidence, they are part of the process and need training. They need training on their tools and techniques as well as training on effective testimony

Many of us have written beautiful requests for training. If you are stuck and need some help, reach out to your friends at LEVA, or NATIA, or the IAI ... or send me a note. We are all here to help.


Benjamin Wright said...

Jim: A new way to promote a good digital chain of custody is to authenticate records (e.g., photo or video) with a voice signature, which helps to show who collected (or manipulated) the evidence, when it was collected (or manipulated), and that it has not changed since collection (or manipulated). What do you think? --Ben

Jim Hoerricks, PhD said...

Your work sounds interesting. The one thing that worries me is the statement, "at present the service is free." What happens to those who used the free service once you begin charging a fee?

Am I understanding your site in that a piece of evidence is bundled with this voice print / hash on your server? This would mean that you would retain the evidence would it not? Would you just pass it through and back to the user to store?

Again, it sounds interesting. I would think that you'd get more traction if you let the user license the software and host it themselves - taking the liability of mishandled or leaked evidence off of your shoulders.

Benjamin Wright said...

Jim: I am very interested in your comments.

The voice signature service aims to provide the user an end-product that is self-explanatory, which means that the user should (at the end) possess an archive that contains everything a third party will need in the future to assess the voice authentication. Thus, the value of the archive does not depend on the continued existence of Voice Signature, its service, its people or its records.

When someone uses the service for free today, obviously Voice Signature is licensing them to use the service and the records it creates (consistent with the purpose of that service and those records). The fact that Voice Signature might in the future charge for new uses of the service does not imply that previous users must retroactively pay for previous uses.

As for storing evidence and records: The service is not an archive storage service. It is the user's prerogative and responsibility to store archives after using the service.

To minimize the records exposed to the service, a user can upload just a hash of the original evidence file, rather than the file itself.

I agree there could be a market for a user-hosted product. Voice Signature has a user-hosted product that is marketed to large enterprises to enable the legal signing of transactions like contracts. There is no reason in principle why a user-hosted product could not be marketed to forensics experts.

Again, I'm interested in hearing your thoughts. --Ben