Digital Enhancement and Transmission of Latent Prints
Who Will Set The Standards?
(This article, submitted by the author, should be read and discussed by everyone concerned with the effects of the digital imaging on the science of fingerprints. The article was simultaneously submitted and printed in many other publications.)
By ERIK BERG
Tacoma Police Department
As digital imaging technology begins to replace traditional film based photography for the recording of latent fingerprints, latent examiners and evidence technicians will begin to take for granted the ability to capture and view a latent fingerprint image in minutes, rather than hours. The ability to enhance a marginal latent or even one that would normally be unidentifiable will continue to advance as computer software technology is applied in a manner similar to the way forensic light sources were applied in the late 80's. The limitations of the human eye will continue to yield to the abilities of the computer to ferret out that elusive latent print that had only minutes earlier been hidden from view. The ability to capture and transmit images by phone line to any destination in the world within minutes will also become a factor as criminals continue to migrate from one region to another. With these benefits come additional questions and responsibilities that must be answered before we as a group can go forward as a coordinated and unified body, building upon the experiences of the past while pushing the future before us.
We must always recognize that a digital image falls under the same legal requirements for chain of custody and originality that film based photography and physical evidence must fulfill. The expert who collects the evidence or identifies the fingerprint must still testify that the evidence presented in court is in the same or in substantially the same condition as when it was collected. In the case of an image it must accurately portray the evidence as he/she first saw it. If either of these conditions cannot be met, the evidence will not be admitted.
The issues of digital image chain of custody and alteration were raised during a recent murder trial in Seattle, WA, where fingerprint evidence from the crime scene was brought to Tacoma for digital enhancement. Procedures developed by the Tacoma Police Department to protect the integrity of the original image, enhancement techniques, and methods used to record the enhancement process were demonstrated at both a Frye hearing and at the trial. The defense had hoped to promote the mystery of the technology and show a lack of professional acceptance and standardized procedure. In this particular case it didn't work, but these issues will come up again in other cases and in other jurisdictions, until they have been answered with uniform procedures and information.
Any prudent procedure must include a provision that protects the original image from any alteration. Any enhancement must be confined to a copy of the original image and any techniques used should be documented and repeatable. Most jurors and many of our colleagues still believe it is possible for a computer to alter a fingerprint by changing the position and relationship of minutia in such a way that the latent will be falsely identified and the alteration will go undetected. This has been and always will be a question of personal and professional integrity. It is not a feature nor is it weakness inherent with digital imaging. Digital image alteration or enhancement is carried out by a computer, at the direction of a person, within mathematical precision. Imaging, like photography, can and has been used for many applications, including falsified images on the covers of tabloid magazines. By adopting procedures that safeguard our original evidence and maintain an unbroken chain of custody, we can demonstrate our professional integrity and preserve a valuable scientific tool.
Issues that need to be addressed by the IAI membership in order to facilitate a smooth transition to digital imaging and provide the membership with enough information to effectively defend the technology in court, include:
1. What definitions and terms are we, as an organization, going to adopt when describing the technology and digital enhancement techniques to colleagues and the legal system?
2. Should there be a minimum acceptable standard for chain of custody procedure, image resolution, file format, and output quality?
3. Should latent images be compressed?
4. Should we impose a standard on the transmission of latent images to ensure there is a baseline compatibility between agencies?
Many of these questions were raised before the IAI's Digital Imaging Committee during the 80th Annual IAI Education Seminar in California last year. The issue of image quality was raised in connection with the Electronic Fingerprint Transmission Specification for the IAFIS Project1, but so far the specification only addresses questions related to ten print images. Latent fingerprint image standards have largely been ignored.
The first issue that needs to be addressed is a glossary of terms. What words do we use to describe the mechanics of digital imaging and the analysis of the resulting images? Digital image processing has been used successfully in other professions, such as advertising, medicine, space exploration, and micro biology for decades. Each of these professions has adopted terms and definitions to describe the application of the technology to their particular field. In some cases these terms and definitions can be carried over directly to forensic identification. One such term is resolution, which describes the amount of discernible detail in a digital image2. To put this concept into perspective, think of a digital image as a matrix of different colored light bulbs. Each light bulb can be varied in both color and intensity. When this matrix is held some distance form the eye, each individual light appears to blend smoothly with its neighbor forming a continuous tone image. Think of each of these light bulbs as a pixel and you will have an understanding of what a pixel is. Spatial resolution, also referred to as spatial density, is a measure of the number of pixels in a digital image3. If you have two images that are exactly the same size and one image contains 250,000 pixels and the other contains 500,000 pixels, the image containing the 500,000 pixels has twice the spatial resolution. It also has twice the information and twice the detail.
Up to this point the definitions are fairly consistent from one profession to the next, but now we'll talk about the concept of pixel density and how it relates to output quality and minimum resolution standards. The graphics profession and many other related fields have adopted a term of dots per inch (dpi) to describe image resolution and, ultimately image quality. Digital capture device manufactures and most of the scientific image processing community use the term samples per inch (spi) to describe image resolution; pixels per inch (ppi) to describe display resolution; and dpi to describe hard copy output quality4. The difference between these terms is subtle, but there is a difference. Over the last several years these terms have been used interchangeably and, in most cases, inaccurately. The definitions that the IAI eventually adopts for these concepts will play a key role in setting accurate standards for both minimum image resolution and minimum output quality. The FBI has determined that, for ten print submission under IAFIS and for Livescan, the minimum acceptable resolution is 500 dpi. If you as a user of these systems do not understand the meaning of these terms, how can you be expected to conform to any kind of standard? How are you going to know whether the standard is good or bad? The rough dpi equivalent of an inked fingerprint is 2000 dpi. How does that compare with the FBI's 500 dpi standard? At least now you should have enough information to ask the right questions. Ordinary latent examiners and forensic professionals must be able to understand and use digital imaging and communicate its use to others if it is going to become an accepted evidence recovery technique. A glossary of digital imaging terms and techniques would go a long way toward the fulfillment of this goal.
Now let's examine the issues of image file formats, latent fingerprint image transmission, and compression. There are currently more than 30 different digital image file formats. An image file format defines the data structures for organizing image data within a computer file. Standard image file formats make it possible for many different digital image processing systems to read and write the same image files by using a commonly understood format3. The Henry classification system could be thought of as a file specification for fingerprint cards. Rules determine where within the filing system a particular group of pattern types are stored, and in what order. A digital image file format is a filing system for pixel information. The type of format determines such things as the order in which pixel data is stored in a file, whether the image is color or black and white, the physical size of the image in inches or centimeters, and whether or not additional data is stored with the pixel information and what that data is. Many formats can store more than one image in a single file. The IAFIS specification is essentially a file format specification for the storage and transmission of ten print images. The specification was developed by the FBI to include provisions for text data, multiple fingerprint images, result codes, minutia data, etc. The cost for implementing this new file format is going to be in the millions of dollars just for software development and integration with existing AFIS systems. Given the volume of ten prints the FBI receives on a daily basis, the need for minimal human intervention during the submission and search process, and the lack of any kind of existing file format capable of supporting the needs of the FBI, a new file format was probably the best solution. Latent fingerprints on the other hand are an entirely different matter. The real issue here is not how do we transmit latent images to the FBI, but how are we going to transmit images to each other? The ability to transmit images between local agencies and the State ID bureau is the real issue facing most of us. Unless you're a member of the Western Identification Network (WIN) or some other regional network, you really don't have the ability to search a latent print against AFIS databases in cities and counties surrounding your jurisdiction, unless you take the time to photograph the latent and mail copies of the latent to individual agencies. Those agencies must then enter and search the latent in their AFIS. This can take weeks. Meanwhile, the suspect is mobile. By next week he could be in one of those neighboring jurisdictions, committing the same crimes. Setting realistic standards and adopting an open image file format would help protect small agencies form costly development requirements that could effectively exclude them form this valuable technology.
The Tagged Image File Format (TIFF) was developed jointly by the Aldus Corporation (now merged with Adobe) and Microsoft and was first released in 1986. The current release is version 6 and was released in 1992. TIFF was originally designed to be used as a standard method for storing black and white images in any operating environment. The TIFF format is perhaps the most versatile and diverse bitmap format in existence.5 TIFF currently supports the storage of both black and white and color images and is supported by most paint, imaging, and desktop publishing programs running under the DOS, Windows, UNIX, and Macintosh operating systems. In addition, Morpho, Printrak, and TRW can all import TIFF images directly into their AFIS systems. NEC is the only AFIS vendor unable to support TIFF, but with the release of their new system 21 AFIS, NEC is making the move to the UNIX operating system. With this move comes the opportunity to create a level of compatibility between not only agencies, but between agencies and every major AFIS vendor. By adopting an open file format such as TIFF, the IAI could put into place a standard that every agency, regardless of size, could afford while ensuring compatibility between all users.
Image compression is normally done as a prelude to either image transport or storage. The benefits of compression include smaller image files, which means less storage space and shorter transmission times when sending files electronically to other sites. Image compression schemes can be divided into two general groups, lossless compression and lossy compression. Lossless image compression preserves the exact data content of the original image, but because it preserves the exact data of the original, only a small savings in storage space requirement are possible. Lossy image compression preserves only a portion of the original data content, which enables a significant savings in storage space requirements. The down side of lossy compression is loss of image quality and in some cases the introduction of artifacts. A poor quality latent print can actually lose minutia when compressed excessively or when certain compression schemes are used.
The questions that must be answered when considering whether or not compression should be used are:
1. Are the benefits of compression worth the loss of image quality and the added potential of having to defend the process in court?
2. With the latest drop in the cost of storing digital images and other computer data, is there really any reason to even consider the compression of images?
The short answer to these questions is no. An uncompressed grayscale image (256 shades), 1 1/2" by 1 1/2� with a resolution of 600 dpi is 792 kilobytes. Using a high speed modem and an ordinary phone line, it is possible to send that image anywhere in the world in less than five minutes. The current cost for hard drive storage is 2 cents per kilobyte. Cost to store this image short term is $1.59. Since the cost for long term archival storage on CD ROM is only .000013 cents per kilobyte, the actual cost for storing this 792 kilobyte image drops from $1.59 to less than 2 cents. (Once the image is transferred from the hard drive to CD, the hard drive space can be reused.) Image compression would further reduce the amount of time it takes to transmit and the cost of storing an image, but given the extremely low costs for the transmission and storage of uncompressed images, it hardly seems worth the potential consequences.
There are many other side issues that could be dredged up and added to the list, but adding any additional issues to the discussion before we have addressed the issues presented here, would serve only to confuse and confound the very people this technology can help the most, latent print examiners. Are we as an organization, dedicated to the advancement of the Forensic Identification Profession, going to wait for some other professional organization or governmental agency to set standards for us, or are we going to impose standards upon ourselves, based upon our needs? Only you, the membership can answer that question. I encourage you to send your comments to Mr. Pat Nolte, Chairman of the IAI Digital Imaging Committee. I've listed his address below. If you have questions or comments you would like to share with me, feel free to give me a call or send an email.
Chairman/IAI Digital Imaging Committee
Anaheim Police Dept.
425 S. Harbor
Anaheim, CA 92805
The author Erik Berg can be reached at:
Tacoma Police Department
930 Tacoma Vs..
Tacoma, WA 98402
1 Federal Bureau of Investigation, Electronic Fingerprint Transmission Specification. Washington, D.C., August 24, 1995.
2 R.C. Gonzales and R.E. Woods, Digital Image Processing. New York : Addison--Wesley, 1993.
3 G.A. Baxes, Digital Image Processing. New York: John Wiley & Sons, 1994.
4 D. Blatner and S. Roth, Real World Scanning and Halftones. Berkeley: Peachpit Press, 1993.
5 J.D. Murray and W. VanRyper, Encyclopedia of Graphic File Formats. Sebastopol, 1994.
This article was printed in �THE PRINT�
Volume 12(4),July/August 1996, pp 6-9
and has been obtained from the online library provided by the
Southern California Association of Fingerprint Officers