Figure 1a
Image of dust shoeprint on chair.
The evidence photographer is faced with a wide range of photographic challenges at crime scenes. Recording crucial evidence in the form of fingerprints, shoeprints, or other impressions is often problematic. These images may be extremely fragile as well as difficult to see. The background color or texture can easily overpower subtle detail. Curved, specular, or uneven surfaces may resist attempts to record all of the detail in one exposure. Lastly, exhibits bearing evidence can often be difficult or impossible to remove from the scene for photography in a controlled environment, without risking damage to the evidence.
In most circumstances, conventional photographic procedures are used successfully to record such evidence discovered at crime scenes. The experienced photographer may employ oblique, tent or coaxial lighting, tilt and swing camera movements, or polarizing and contrast filters to obtain a suitable recording.
Digital evidence enhancement techniques, such as Fast Fourier Transform and Image Subtraction, have been described previously in detail. It is not the purpose of this study to reiterate those descriptions or to report case results, but rather to recommend specific procedures for the maximum recovery of evidence when analog techniques are suspected to fall short. The successful application of computer technology depends on the technician recognizing the potential of these procedures and following the required necessary steps in recording images at the scene.
Four situations, based on the collective case experience of the authors, were re-created to explore the effectiveness of digital techniques, the application of which depends on recordings taken at the scene in specific order and procedure. Although the images acquired in this study were transferred to a desktop computer for processing, the same results could have been obtained on a laptop computer at the scene. Resulting digitized fingerprints can be forwarded for AFIS search with minimum delay.
All images were captured with a Nikon D-1 digital camera. The images were captured as 24 bit RGB images in a TIF format (lossless) and were 7.56 MB in size. The resulting images were processed using Image-Pro Plus, Version 4.1. The image subtractions were completed while the images were in TIF format. The images were lastly converted to 8-bit gray scale JPEG images for ease of transmission and publication.
Image subtractions were completed with the color images. It is significant to note that if the images are converted to gray scale prior to the subtraction, an entirely different and greatly inferior result may be obtained. It is also significant that other examples, prepared in identical fashion, were unsuccessful in revealing additional detail.
A dust shoeprint was deposited on the cloth surface of a chair. Similar impressions have also been discovered on rugs and other fabric surfaces at crime scenes. The fabric caused obstruction of the impression both in surface discontinuity and in color pattern. This type of evidence, routinely encountered at break-and-enter scenes, is extremely fragile and may be lost or degraded by the removal, or even jarring movement, of the chair. In actual case applications, it is strongly recommended that the impression be recorded by conventional photography to the extent possible, before the digital recording process is started.
A scale was placed beside the impression. A Nikon D-1 was placed on a tripod over the chair, and an image was recorded (Figure 1a). Without changing the position of the camera, the lighting, or the chair, the scale was removed, and the impression was vacuumed, with care taken to avoid movement of the chair or fabric. A second image, in register to the first, was taken.
The images were later downloaded to a desktop computer, and background subtraction was performed (Figure 1b). The image was inverted and was adjusted for contrast (Figure 1c).
Fingerprints were placed on patterned wallpaper and treated with ninhydrin, in re-creation of an actual homicide scene examined by two of the authors. In such situations, the wallpaper is frequently the non-strippable variety, and it must be examined in situ.
A scale was placed beside the resulting impression and the impression was recorded using the Nikon D-1 (Figure 2a). Without altering the relationship of the camera to the image plane, the scale was removed and the fingerprint was erased by the application of a mild solution of bleach (20%) with a cotton swab. After allowing the surface to dry, a second image was captured.
As in the previous situation, this second image was subtracted from the first (Figure 2b). The resulting image was adjusted for contrast and was processed with Fast Fourier Transform to improve clarity of ridge detail and to suppress background interference (Figure 2c).
A fingerprint was deposited in grime on patterned wallpaper to emulate visible dirt impressions where chemical development is not required or possible. A digital recording of the impression and scale was obtained with the Nikon D-1 (Figure 3a). The fingerprint was then erased with a damp swab, preserving the register of camera to subject. When the surface had dried, the scale was removed and a second image was obtained, which was subsequently subtracted from the first image (Figure 3b). The resulting image was adjusted for contrast and was processed with Fast Fourier Transform to improve ridge clarity (Figure 3c).
A fingerprint was developed on patterned wallpaper with ninhydrin, similar to scenario two (Figure 4a). The detail was observed to be distinct in some areas, but was interrupted by the pattern of the wallpaper. The ninhydrin fingerprint was erased as previously described in scenario two, and a second image was recorded. After image subtraction (Figure 4b), detail was revealed in the obstructed area which resulted in a much larger area of continuous ridge detail. The image was adjusted for contrast (Figure 4c).
*From the Journal of Forensic Identification Vol. 52, No. 6, November/December 2002
The Official Publication of the International Association for Identification
"Reproduction of the Journal of Forensic Identification, in whole or in part, for noncommercial, educational use is permitted provided proper citation of the source is noted. Reproduction for any other use is prohibited without prior written permission. Requests for permission may be addressed to the editor (of the Journal of Forensic Identification -- jfieditor@theiai.org)."
Return to the CSI Articles Page