Why Accuracy Of Forensic Latent Fingerprint Matching Still Depends On Human Expertise?

Human identification on the basis of fingerprint has been closely associated with forensics. From early usage in criminal investigations to modern AFIS (Automated Fingerprint Identification Systems) setups, traces of human friction ridges found at crime scenes have been presented and accepted as valid evidence in judicial proceedings. Modern applications, however, are not limited to criminal investigations or law enforcement applications. Fingerprint recognition has made its way to civil identification as well. However, despite that ground breaking success of fingerprint identification in civil identification, forensics and law enforcement has kept its relevance intact and technological advancement has taken it to the next level.

In the subsequent sections we will discover why fingerprint identification is the most prominent method at forensics and law enforcement; how latent fingerprints are captured and matched with reference prints, and why AFIS is yet to fully automate latent fingerprint identification.

When it comes to criminal investigations and forensics, fingerprint recognition is the oldest and most used evidence around the world. Though various other methods of biometric identification have emerged with advancement of the technology, fingerprint identification still remains the most prominent identification method of crime scene investigations. A fingerprint is the impression made by the papillary ridges on the ends of the fingers and thumbs. Secretion from pores in this area contains sweat and oil and when touched with any surface, it leaves the pattern of friction ridges on it. Later we will cover how these prints becomes the most important evidence in crime scene investigations.

• Fingerprints afford an infallible means of personal identification, because the ridge arrangement on every finger of every human being is unique and does not alter with growth or age.
• People keep touching different surfaces all the time and leave their fingerprints there. Fingerprints are the most probable human identification trace to be found at crime scenes.
• Probability of finding two people with the identical fingerprints is extremely low.
• In over a century’s usage, forensic science has well accommodated itself with fingerprint recognition and there are standard operating procedures for most processes. From capturing latent prints to matching them with existing ones.
• Modern fingerprint identification systems are more advanced, cheaper and technologically assisted than any other biometric recognition method.

A latent fingerprint is the two-dimensional reproduction of the friction ridges of the finger on an object by means of perspiration, oils, or other contaminants that coat the surface of the ridges when a finger touches a surface. Since fingerprint is an individual characteristic, latent fingerprints unintentionally left at crime scenes can help law enforcement agencies to associate them with a known identity and solve the crime. These fingerprints are processed with different forensic fingerprint development methods like fingerprint powders, chemicals or alternate light source.

Whenever a criminal incidence occurs, forensic experts are among first people to call, so that they can search and collect human prints before they are contaminated. Forensic experts have specialized methods to collect prints from different surface and are trained to deal with different situations. Latent prints are formed when the body’s natural oils and sweat on the skin are deposited onto another surface. Latent prints can be found on a variety of surfaces; however, they are not readily visible and detection often requires the use of fingerprint powders, chemical reagents or alternate light sources. Latent fingerprints can be very crucial evidence that can solve a crime if collected skilfully, or complicate it if missed.

Fingerprints found at a crime scene, may potentially belong to people involved in a crime or present at the time of incident. Since such fingerprints can be there but may not be visible, they are called latent prints. Forensic experts are called in to look for and collect these prints, which make use of sophisticated methods of making these hidden prints visible and collect them for analysis and matching with other prints.

Dusting is the oldest fingerprint development method to discover and collect latent fingerprints. It can provide excellent results if done skilfully, however, it can also potentially damage or distort a latent fingerprint if performed imprecisely. Dusting can help develop latent fingerprints from smooth or non-porous surfaces using a fingerprint powder like aluminium flake, black granular, black morganatic, etc. However, use of fingerprint powder can contaminate the surface by leaving it unusable for other method, if dusting fails to recover fingerprints. Choosing right method for developing latent fingerprints requires careful examination of the area and surface where there is a possibility of latent prints or additional information.

The surface can also be examined with super glue (cyanoacrylate) or an alternative light source. Alternative light source emits a particular spectrum of wavelength of light from using laser or LED, which can help latent prints appear more clearly. A forensic expert may also choose to process a surface with latent prints with cyanoacrylate (superglue) before applying fingerprint powder. In this process, the surface is fumed with cyanoacrylate. If the surface has any latent print, the cyanoacrylate fumes stick to it, making latent print visible.

Dusting renders unusable on a porous surface (e.g. paper), so latent prints are captured with chemical process. In this process, the surface is treated with chemical like ninhydrin to make latent fingerprints documentable. Ninhydrin reacts with components found in fingerprints and the print turns purple. A different chemical called DFO (1, 2-diazafluoren-9-one) also used to get prints from porous surface, which causes fingerprints to glow, when illuminated with blue-green light.

Since the method of developing fingerprint largely depends on the surface they are being collected from, there are different methods that have been developed to collect latent prints from different sorts of surfaces. When it comes to collect fingerprints from more difficult surfaces like clothes, skin, textures, curves, etc., special techniques are leveraged. Since some methods can render the surface with latent prints unusable for others, first methods with least side effect or destructive tendency are used. Once latent fingerprints are successfully collected using any of the suitable method, the identification process starts.

Technological advancement in fingerprint recognition technology has greatly improved the identification process. Law enforcement agencies now make use of AFIS (Automatic Fingerprint Identification System) to automate the identification process, however, in case of latent fingerprint identification, manual intervention is required to extract fingerprint features and verification.

Once latent prints are successfully captured, they are transferred to a fingerprint expert to analyse their viability for the further comparison stage. If not enough features are present or are of inadequate quality, prints are not processed further and reported as inadequate. If prints are found to be quantitatively and qualitatively good, they are taken to do a comparison against other fingerprints. To be eligible for being accepted for further processing, the latent fingerprint should have sufficient ridge information.

When a latent fingerprint is found to be good enough to be processed further, it is taken to the next step of fingerprint identification process: comparison. This is the step where collected latent print is matched against a reference print to discover their similarity or dissimilarity. Latent can be in a less than ideal condition, and finding similarities (or dissimilarities) requires careful examination. Law enforcement agencies use AFIS systems to find a match of rolled or plain fingerprint images. These images generally have great amount of details and can be processed without the need of human intervention. However, latent prints have their own set of challenges, that make automatic identification systems struggle and results can be highly unreliable or inconclusive.

In an ideal scenario, AFIS should be able to extract features from query fingerprints (latent) and match them with a gallery database (rolled, plain, or even latent images) to obtain a set of possible “hits” with high confidence so that no human intervention is required. However, this is generally not the case, and automatic identification of latent prints require great amount of human intervention during feature extraction from a latent, e.g., orienting the fingerprint, marking the region of interest, etc. The system then outputs a short list of candidates that need to be examined by a latent examiner to determine if any of these fingerprint comparisons is a match.

Evaluation is a conclusive step in which the fingerprint pair (latent and the reference print) is determined to be a match, no match or to be inconclusive. Only a well performed comparison can lead to an efficient evaluation. If there are no unexplained differences between latent and the reference print, and the individual characteristics are in agreement, the pair is concluded to be a match. If the comparison results otherwise, is a no match. Sometimes, neither of these conclusions is possible as there may be not enough ridge information present in the latent print to make a comparison. In such cases “inconclusive” is the result to be returned.

Verification is done by an examiner other than the first examiner who came out with the initial conclusion. During the verification, the whole process is independently repeated by a second expert and final conclusion depends on the agreement of both the experts. If the examiners are in disagreement, further analysis can be required. The verification process affirms the conclusion and reduces the possibility of human errors.

AFIS has greatly reduced human efforts required for fingerprint identification. Today’s AFIS systems are backed by powerful hardware and computing resources, which can process fingerprint matching operations faster than ever. However, when it comes to latent print identification, possibility to automatic identification depends on a lot of things. Conducting AFIS latent print searches may not be entirely possible without human intervention, but the amount of manual efforts can be greatly reduced with the help of AFIS. Searching latent prints from crime scenes against large fingerprint and palm print repositories can be impossible with all manual efforts. AFIS technology has been extremely beneficial in this regard, running a search query against large numbers of fingerprint or palm print records.

• Qualitative factors such as clarity are generally lower in latent prints, making AFIS systems unable to process them.
• Latent prints are generally of lesser size than what an AFIS system can accurately process. It lowers the quantitative factor.
• The orientation of the print is generally less certain
• The finger position of a latent print is also mostly uncertain.
• Presence of the background noise due to the process used for lifting latent prints from a surface.
• The latent print, which is considered to be a fingerprint, might actually be originating from some portion of friction ridge area, such as the palm. (Matching which against a database of fingerprints will be waste of time and resources).

Due to these difficulties associated with latent fingerprints, human intervention is required, specially when prints are quantitatively or qualitatively inadequate. Despite the challenges, AFIS can always make things faster if a fingerprint expert can engage side by side, way faster than all manual process.

Fingerprint identification has been associated with law enforcement and forensic applications for more than 100 years. Despite the advancement of biometric technology and rise in its civil applications, fingerprint identification remains the most prominent method in crime scene investigations. Searching and collecting latent prints are the activities done by forensic experts whenever a crime takes place. These prints then go through analysis and comparison phases to make a conclusion. The final conclusion is again verified by another expert to affirm its validity.

Present day fingerprint forensics greatly depends on AFIS, however, when it comes to latent prints, expertise of AFIS fall short due to varied number of challenges present in latent prints. To overcome these challenges, a forensic expert has to intervene to keep identification process free from errors. Accuracy of forensic latent fingerprint matching greatly depends on human intervention, which is expected to remain a constant requirement, until AFIS systems become as good as forensic experts!