Digital Imaging for Fingerprint Enhancement (DIFE) involves the conversion of a picture or image of a print into a digital file, usually by the use of scanners and digital cameras. The scientific process for this is by use of a grid of detectors known as a charge coupled device (CCD) that detect light patterns and convert them to an electrical signal. The detectors each correspond to the pixels on a digital image; each pixel takes on a particular shade of grey according to the quantity of light the sensor it is representing detects.
The more sensors used in the CCD, the more pixels can be created and therefore the more accurate and detailed is the representing image. The digital file created can then be enhanced in several ways; an interfering background can be removed from the image, the contrast between the print and the background can be increased and the intensity difference between each pixel can be increased in order to sharpen up the image. Digital imaging also allows computer-aided image matching in which the software can search for similar prints for subsequent analysis by the fingerprint expert.
The CCD is an electronic imaging system where the process of light capture, is performed by photo diodes. Here each incident photon frees up an electron, which can be stored for a subsequent readout, the charge is then temporarily transferred via a gate from the photo diode to the charge coupled device. The CCD acts like a bucket holding the electrons. The CCD's can be cascaded into arrays in which the electrons can be efficiently moved from one cell to another over long distances to a final readout destination. The reading, which is taken from each array, is directly proportional to the light intensity at that point in the captured image.
So just as in conventional photography where silver halide crystals record the level of light intensity, the photo diode found in a digital system does exactly the same job. The intensity value from the CCD is subsequently stored, as a value, which relates to the grey scale value at that point within the image. The subsequent image is finally composed of a series of square picture elements, pixels, each with it's own numerical value. We now have an image that is essentially composed of a series of numbers in square blocks, rather than silver halide crystals.
Most digital cameras capture images using an 8-pixel depth. For fingerprint photography it is preferable to use a 12 bit pixel depth which will give 4096 levels of grey scale rather than just 256 with the 8 pixel depth. This greater level of grey scaling means that the fingerprints can be seen in far better detail than before showing greater detail for analysis, especially if needed to be used in evidence to show possible suspects. One of the greatest benefits of digital technology is the availability of image enhancement tools.
When used in a controlled environment these facilities offer considerable benefits to the crime examiner. As the photograph is converted to a series of numerical values it is then possible to use mathematical algorithms to change the values and so improve the quality of the final images. Different types of digital filters can be used to enhance the fingerprints, however one of the most simplest and most common ones is the Kernel Mask. This mask has a two dimensional array of numbers that can be used to alter an image in order to achieve a desired effect of smoothening or sharpening.
The numerical values in the filter have different effects on the image. They can also be used for point, line or edge detection for image segmentation and analysis. To show how effective digital imaging is for analyzing fingerprints, the pictures below show the same fingerprint but clearly show that one is far more identifiable than the other. The fingerprint shown was developed with ninhydrin, but clearly shows a great smudge mark in the original picture on the left. The picture on the right however shows the ridges and furrows of the print far better and this is because of the use of a mask.
The larger the filter is, the greater effect on the image it will become. As with the fingerprint shown below, 3x3, 5x5 and 7x7 sharpening filters were used in succession followed by a 3x3 median filter for smoothing. This clearly helped to achieve greater clarity. In order to achieve the desired effect on the fingerprint using mask operations, a piece of computer software called Image Pro Plus is used. This scientific image analysis and enhancement software is very adaptable and has an extensive support network.
It gives the forensic photographer access to a wide diversity of exceptional image enhancement facilities. It works with 12-bit, 16-bit and floating-point images and has a very effective and easy to use Fourier transform algorithm. It is able to show you the direction of the wavelengths in the diffraction pattern, and the image at the same time, and can both boost and cut any selected frequency spikes. Forensic users can select from a range of algorithms developed, which allows them to see fine detail such as ridges or loops in their final fingerprint image.
A 'depth map' showing which parts of each fingerprint image have been used to make the final one are also generated and displayed. As the mouse cursor is moved over the image, each contributing fingerprint image is reported in real-time. To protect the original fingerprint evidence, these systems have built in software, which tags the image and preserves the integrity of the original fingerprint image. This built-in validation ensures that suspects cannot state in court that the image has been falsely enhanced and makes it difficult to dispute the digital fingerprint evidence.
The numerical values or taps in the filter have different effects on the image. If the filter values are positive the resulting image will contain mainly low frequencies therefore generating a smoother softer image. Filters with these values are called low pass filters. The larger the filter, the flatter the filter weights, the smoother the resulting image. To sharpen the image you need to use different filter values. If you place a positive value at the centre of the filter and put some negative taps in the adjacent spaces when you use the filter it will sharpen the image.
This is a high pass filtering operation. The larger the contribution of the negative taps, the sharper the resulting image. The more an image is sharpened the more the noise in the image will be amplified, but used with care this type of filter can be very effective especially on smudged Ninhydrin fingerprint impressions as shown below in the pictures. The high pass filtering operation is evident in the picture on the left whilst the one on the right shows the original photographed image taken from the scene of crime.
The fingerprint on the left has been enhanced with the use of a Flatten filter, the fingerprint now contains enough information for a possible identification. Using the Image enhancement software, it is possible to create a different number of desired effects on the fingerprint dependant on what it is the investigator is looking for. For instance, it is possible to change the background colour of the fingerprint to that of the fingerprint itself and change the fingerprint colour to that of the backgrounds colour.
This can help identify ridges and loops more easily when the background is made either darker or lighter from different grey scale values. These techniques used to enhance fingerprints have so vastly improved since the introduction of digital applications. The use of digital fingerprints as evidence in courts is now classed as admissible because they are regarded as being accurate as well as a primary source of evidence. Many convictions of criminals have occurred due to the success of this technique and it is improving even further especially with the introduction of new software programs.
However if possible the technique could be improved further if it were possible to make the images three-dimensional. This could mean that the fingerprint could be looked at and enhanced from all directions making clarity and detail even greater. This would prove extremely useful in applying digital imaging to fingerprints which are not recorded on flat surfaces where just photography alone may prove to be difficult, especially if the fingerprint is on a curved surface. Overall Digital Imaging has changed the way we portray evidence of fingerprints making it much more accurate and will continue to do so in the future.