I was asked recently to write an article on Fingerprints and while doing research I came across the story of Victor Reyes, a Florida man who was put on trial for murder based on a fingerprint found at the scene and ultimately linked to him. What caught my eye was the fact that the latent print used against Reyes was digitized by police and “enhanced” with a couple of image manipulation software programs. Adobe PhotoShop is perhaps the most well-known imaging software in the world; the other product, MoreHits, is less well known but has tremendous presence in the law enforcement latent print examiner community.

Use of such image enhancing products raises tough questions about chain of custody issues and preservation of evidence. Police using these products swear that they’re not altering the image in any significant way, just “enhancing” it to make it more “readable”. But how can we know this? Moreover, once it’s accepted that they’ve altered the image in the first place, how can we know how far they went with their alteration? What record is there? What kind of audit trail is there detailing what transformations the image has taken while in police custody?

The digital forensic fingerprint community is trying to address these problems with things like log files that track changes made and digitally encrypted original images that are preserved while enhancements are made to copies. These changes are welcome, but I don’t believe they’re enough. The problem is that latent fingerprint identification is inherently a subjective process where human judgment is called upon to compare minutiae in one set of ridgelines to another, and the degrees of difference are sometimes very subtle. After all, the top FBI experts in the Brandon Mayfield case declared a “100% match” between the latent recovered in Madrid, and Mayfield’s prints, and even his own defense expert agreed that it was a 100% match. And they were all wrong. With this much subjectivity affecting the process, it’s not clear what good it will do to preserve the original, unedited images without a better understanding of why examiners come up with identifications in the first place. Adding “enhancements” to the mix just clouds the issues further by introducing new ways to affect the subjective judgment without any understanding of how. What effect will we see from these “small enhancements”, and which way will they cut? In favor of more false positives? Fewer false positives? No one knows.

As for Reyes? He was aquitted. When asked about the verdict, the jury foreman (who was himself a computer imaging software engineer) expressed the jury’s unanimous skepticism of the technology employed by police. There’s that CSI effect again.

Read about the Reyes case here.

RP

This entry was posted on Tuesday, June 27th, 2006 at 10:34 pm and is filed under Software, Junk Science, Fingerprints. You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.