HDR Latent/Evidence Photography

[Boyd Baumgartner]

Has anyone done any HDR photomerging with their latent prints in photoshop? If so, are you finding any processing techniques or substrate types that standout as more successful than others.

I've just started experimenting and have only done CA + Basic Yellow on a Rockstar can, so if you've got some recommendations, I'd like to hear them. I'm also going to start experimenting with the Infrared HDR as well, so if you've got experience/suggestions with that, I'm all ears.

thanks,

Boyd

[George Reis]

Hi Boyd:

I haven't done much with this, but do think it's a great idea. Will you post some sample images? You may also try placing the multiple frames into an image stack and experimenting with the various stack modes. I look forward to hearing more as you do more experimenting!

George

[bayareacriminalist]

Boyd and George, HDR is truly a fascinating area that may be the future of latent print enhancement. Check out the new JFI issue Nov/Dec 2008 for an informative article (with images) on HDR fingerprint images and Photoshop.

Good luck!

Vincent P. Deitchman, CLPE
Criminalist
Santa Clara County DA Crime Laboratory
1557 Berger Drive, Suite B-2
San Jose, CA 95112
(408) 918-2947
(408) 298-7501 FAX

[George Reis]

Are the images from that article available for download?

George

[mdavis]

The merging functions in Adobe Photoshop CS3 and CS4 are quite effective beginning with the basic photomerge, to HDR, and in CS4 to the latest vertical stack merging for progressive depth of field. However, we are beginning to skirt the restrictions of limiting enhancement to darkroom equivalent procedures (I believe a SWGFAST recommendation) when we begin to use computer alogrithm manipulation (the source code of which is proprietary and unavailable for defense scrutiny) to enhance photos for analysis purposes. Additionally, while the HDR article was interesting, there are always multiple methods available within Photoshop for doing the same things, and the judicious use of the 10 (or 13) available channels with the Calculations function is more easily defended in terms of darkroom-related terminology of filtering, lighting and exposure. I would be very cautious before presenting any HDR-manipulated (er, enhanced) images in a courtroom situation.

[Gerald Clough]

So far as I can find, SWGFAST does not recommend avoiding any but what SGTIT calls "traditional enhancement techniques." It does, as it should, call for preservation of original and working files and documentation such that the processes can be replicated and evaluated. SWGIT goes into considerable detail of the natures, benefits, and hazards of various enhancement, restoration, and compression processes. SWGIT also states:

Processed images are used for many purposes by the forensic science community. They can yield information not readily apparent in the original image, which can assist an expert in drawing a conclusion that might not otherwise be reached.

And they frankly include enhancement in processed images.

I think this accurately expresses the intent. Available techniques that enhance the value of evidence cannot be rejected simply because the software code is unavailable. One might also attack darkroom techniques used to produce images without an accompanying highly technical and product-specific explanation of the processes. As SWGIT notes, all of the techniques addressed as enhancement have roots in the darkroom silver processes. In general, imaging software merely provides a more efficient way to accomplish the same things. Some are direct analogues. The unsharp mask's odd name reflects the fact that it does exactly what we would did with a physical mask. It only seems nifty because not many photographers did it until they could do it digitally. We could do it step by step in a digital environment and indeed had to before it was reduced to a predefined process. I could create the mask and apply it digitally, but that would be pretty silly.

Similarly, the combining of bracketed exposures to display high dynamic range images dates to the 1850's (limited film response drove the effort) and notably from the 1930's. Some enhancements of course have no silver analogue, since they involve operations on pixels relative to other pixels. But it can be hard to make overall true statement about non-traditional techniques. Noise reduction, for instance. Actual noise doesn't occur in film. (Grain is something else.) But even digital noise or sensor dust spotting can be addressed with traditional methods with more or less trouble and effort, and a lot of what's done digitally is exactly what would be done in the darkroom. And while, admittedly, non-linear contrast controls, curve adjustment, is slicker in digital, something similar could be applied in the darkroom, if you wanted to make a heavy investment in time and experimentation. Some things - pattern noise and demoireing and halftone removal - would be beyond all but the most adept darkroom artist, who would probably go mad in the process.

The point is that, if the original image, working images, and documentation of the specific processes applied and the workflow are maintained, any advocate can fairly argue the toss. And latent print imaging is perhaps the most straightforward application. We are not often concerned with color rendition, and our images are normally captured under controlled conditions, and such things as blur reduction are not needed. And, since we are accustomed to dealing with distortions, we wouldn't often feel the need to do geometric adjustments, except to explain the distortion. These things are, of course, much bigger issues in some crime scene work. And for perspective, we don't feel obliged to come to court armed with knowledge of chemical interactions with particular substrates to defend against the argument that this or that detail is some kind of migration of developed matrix through the paper that only looks like it's ridge.

We do have to be ready to talk about what was done and why it doesn't create or destroy. The digital operator should be trained and available to testify. (Especially important today, since that could well be a reasonable extension of the principle expressed in the recent SCOTUS ruling on lab analysts being available in person.) The reality is that our most likely task is to defend against an implication that "you could have done anything with Photoshop." But juries do readily recognize that digital enhancement is in routine forensic use. (They have probably been led to believe you can do legitimate miracles.) And you have SWGFAST, SWGIT, and publication of articles like that in the JFI to back you up. And the availability of code is really not an issue. In the unlikely event it was needed, expert testimony on the nature of the processes and what it can or can't do is readily available and applicable to whatever tool is used, PS, PSP, Gimp (including it and open-source plug-ins for which code is openly available), or whatever. But I will agree that no one should walk in with an HDR image product without being thoroughly familiar with how it works and with a decent prepared talk explaining it and without having prepared the attorney they're working with.