3D Fingerprints? Is it time?

[Shane Turnidge]

Hi folks, here's a brief article on some of the rationale behind moving towards creating and maintaining known three-dimensional hand records from the current two-dimensional live scan records.
I'd love to know your thoughts.

Shane Turnidge

It’s time for a fingerprint revolution!
By: Shane Turnidge

Here in 2017 you’d think the last thing we’d be talking about is a revolution in forensic biometrics. But I believe that is exactly what is needed.
For the best part of three decades we have been committed to live scan technology to acquire fingerprints and palm prints from known individuals, and for three decades we’ve been apologists for the limitations of these records. The limitations of these process are well known. I’ll touch on a few issues below.

We know the most significant issue with live scan records are the artifacts (false level two features) that can be created during the image acquisition process. These artifacts differ somewhat from the artifacts we used to create with ink and paper because the smudging and tonal shifts that helped us identify and resolve these artifacts is removed by image enhancement processes in the live scan software. This is done to make the final image more visually appealing.

Friction ridge artifacts are created during the transition between the three-dimensional object (the hand or fingers) and a two-dimensional surface. These artifacts usually appear near the tips but not always. They can occur where you least expect them, and when they do they can easily lead to a false negative evaluation, especially if the latent is not robust.

We also know of several other problems with certain live scan systems. These include issues such as feathering, ghosting, stitching, and over exaggerated friction ridge deviation. Perhaps the most surprising limitation is, that because of the method used in acquiring images we only end up with half of the friction ridge detail available on the fingers and thumbs. This is due to a technique known as, “rock and roll”. The technique requires a gentle rocking of the finger to activate the sensor. The technician then fills the screen as best as they can, to capture the fingerprint. I have spent time observing this process in a lock-up and noticed that the fingers and thumbs were only being rotated 45° in each direction. Conversely, when we acquired fingers and thumbs using the ink and paper method we rolled the prints from nail edge to nail edge. Or 90° each way. The additional 45° degrees x2 of information becomes extremely valuable in resolving ordinary latent prints from things like beer bottles and cans where we typically see the very edge of index fingerprints and thumbprints.

So, what is the solution?

I believe the solution can be found by moving away from two dimensional known records. We need to create known records in three dimensions. Hands and digits are three dimensional objects and should be recorded in three dimensions. The impressions they create are two dimensional (most of the time). I’m certain, in the computer age that we can render a three-dimensional object into a two-dimensional image, we’ve done it with maps for years. We can use the computer to render the portion of the friction skin we need to resolve the latent print, while maintaining the integrity of the known record.

By capturing hands in three dimensions we will have access to all the friction ridge detail available. We will eliminate artifacting. We will capture all friction ridge detail including the detail sometimes not acquired due to deposition pressure or natural physical depressions. We will capture all the friction ridge detail from nail edge to nail edge including tips. We will capture the friction ridge detail on the phalangeal joints often unrecorded in flat impressions. We will acquire all the friction ridge detail in three dimensions including detail in the furrows that is sometimes not captured unless enough deposition pressure is applied. This becomes very relevant when comparing photos or videos of hands and fingers.

By adapting to using three-dimensional imaging of known hands we will maximize the opportunities to resolve almost all the very important latent finger and palm print evidence collected by our law enforcement professionals.

[Dr. Borracho]

I clicked on the link and got a message that the site is not available. I closed "CLPEX" and retried several times, but with the same result. You made me curious, but I can't get there to read the story.

[Shane Turnidge]

Apparently the link was dependent on a LinkedIn account...
I'll put that one under future considerations.

Shane Turnidge

[Steve Everist]

We know the most significant issue with live scan records are the artifacts (false level two features) that can be created during the image acquisition process. These artifacts differ somewhat from the artifacts we used to create with ink and paper because the smudging and tonal shifts that helped us identify and resolve these artifacts is removed by image enhancement processes in the live scan software. This is done to make the final image more visually appealing.

We also know of several other problems with certain live scan systems. These include issues such as feathering, ghosting, stitching, and over exaggerated friction ridge deviation.

By adapting to using three-dimensional imaging of known hands we will maximize the opportunities to resolve almost all the very important latent finger and palm print evidence collected by our law enforcement professionals.

I've mashed portions together in the quote to primarily ask one question - how have we resolved these issues by going to 3D capture?

These problems seem to be inherent to computers being used for capture/processing/storage and not the physical act of taking the print (as is the case with the unquoted problems). If 3D capture requires a computer just as the current Livescan does, would you not still get the same artifacts that "can be created during the image acquisition process?"

And what problems will be inherent to the 3D process that aren't currently?

It seems like there could be several starting with file size/type/storage to how will they be visualized for comparisons on-screen?

Will there be special software required to do comparisons of the 3D images to the 2D latent print evidence?

Will all of the current Livescan machines in use have to be scrapped?

Since latent prints are recovered in 2D (via lifts or photography), what's the benefit of a 3D electronic capture other than getting areas of detail not currently recorded via Livescan? It seems like it would suffer from the same problems you say Livescan suffers from as a result of being a method of electronic capture.

The solution, to me, based on the article seems to be more that we need to take inked MCP's instead of go 3D electronic.

[Dr. Borracho]

Since latent prints are recovered in 2D (via lifts or photography), what's the benefit of a 3D electronic capture other than getting areas of detail not currently recorded via Livescan?

Thinking creatively here, I could imagine software that would allow you to extract from a 3D model what a latent would look like with light pressure versus heavy pressure, or even sideways pressure where the ridges tilt to maybe connect a ridge ending to an adjacent ridge to mimic a bifurcation. With a borderline latent, that little extra adjustment may give you a nudge to either identify or exclude.

Just brainstorming here, Boss . . . .

[Shane Turnidge]

Hi everyone,
Ok Steve...

These problems seem to be inherent to computers being used for capture/processing/storage and not the physical act of taking the print (as is the case with the unquoted problems). If 3D capture requires a computer just as the current Livescan does, would you not still get the same artifacts that "can be created during the image acquisition process?"

The problem of artifacting in prints is associated to the physical taking of prints. Artifacting also occurs in latent prints as well as inked impressions. The difference being that with live scan impressions the evidence required to resolve the artifact is usually destroyed in the image cleansing process.
I think we all can agree that when a skilled print taker uses a live scan device we get reasonable quality. When an unskilled practitioner takes prints and allows the person being printed to create unusual movements during the acquisition process, artifacts become likely.

FWIW i do not think we will see the same type of artifacts we currently see when a computer renders a 2D image from the known 3D image. I liken it to a stamp. Which would you trust more, the stamp or the impression created by the stamp?

And what problems will be inherent to the 3D process that aren't currently?

I don't know. I do believe that the benefits of going to 3D known records far outweigh the potential cons.

It seems like there could be several starting with file size/type/storage to how will they be visualized for comparisons on-screen?

Will there be special software required to do comparisons of the 3D images to the 2D latent print evidence?

File size could be an issue, but as we've seen in recent times, storage is becoming cheaper. I would expect that new software would be developed to handle the comparisons and the renderings. I've worked with some engineers from several AFIS/Live Scan software vendors and I find them to be very good at what they do. We just need to point them in the right direction from time to time.

Will all of the current Livescan machines in use have to be scrapped?

I imagine a transition would take place, but the benefits would be significant.

Since latent prints are recovered in 2D (via lifts or photography), what's the benefit of a 3D electronic capture other than getting areas of detail not currently recorded via Livescan? It seems like it would suffer from the same problems you say Livescan suffers from as a result of being a method of electronic capture.

The solution, to me, based on the article seems to be more that we need to take inked MCP's instead of go 3D electronic.

If you've seen the contactless print systems being offered you might get some insight into how 3D images could be achieved. Maybe someone will come up with a better solution. Part of the problem we currently face is that anyone and everyone is creating our known records whether they are trained or not. We need a system where training is minimal and where the system does the work - quickly. Major Crime Prints are awesome but not necessarily comprehensive. The time required to get MCP's from people with the current systems would end up discouraging the practice.

As much as I am a fan of inked impressions I don't think that going back helps us in the computer age. It's time computers worked for us and not the other way around. Will there be growing pains? I expect yes, but again, the benefits will far outweigh the consequences and I think that after 30 years or more of live scan technology we all realize the limitations of that technology.

Shane Turnidge

[C. Coppock]

As we do recover many impressions from 3D surfaces, any additional information for our comparative potential would likely prove beneficial. This question was raised in my Washington State Daubert hearing back in 2006 or so.

Here is a small relevant article from my blog to get one thinking: Two Dimensional Fingerprints in a Three Dimensional Whorled.

https://fingerprintindividualization.bl ... print.html