This paper examines the practical applications of fingerprint classification in forensic science, covering the three fundamental principles of fingerprint individuality, the major pattern types (arch, loop, and whorl), and the anatomic features used in identification. It details how automated algorithms process ridge data and minutiae to match prints against large databases, including the FBI's repository of approximately 70 million records. The paper also describes feature extraction techniques, classification accuracy for five- and four-class systems, and the role of pattern matching as an alternative to minutiae-based approaches. Future identification technologies such as DNA fingerprinting and iris detection are briefly considered as potential successors to traditional fingerprinting.
The paper effectively uses enumeration and operational definitions to organize complex technical content. By listing and defining each fingerprint pattern type and anatomic criterion with specific measurement thresholds (e.g., ridge length greater or less than 3mm), it demonstrates how scientific writing can present classification systems with precision and clarity.
The paper opens with a brief abstract-style introduction, then establishes the three core principles of fingerprint science before tracing its history. Subsequent sections cover pattern types (arch, loop, whorl), anatomic criteria with operational definitions, feature extraction and algorithm design, and classification accuracy. The conclusion synthesizes the material and gestures toward future technologies. This deductive structure — broad principles first, technical specifics second — is well suited to a scientific overview paper.
Large volumes of fingerprints are collected and stored every day for use in a wide range of applications including forensics, access control, and driver's license registration. Automatic recognition of people based on fingerprints requires that the input fingerprint be matched against a large number of fingerprints in a database. For example, the Federal Bureau of Investigation database contains approximately 70 million fingerprints (Azoury et al., 2004). To reduce search time and computational complexity, it is desirable to classify these fingerprints in an accurate and consistent manner so that the input fingerprint need only be matched against a subset of the fingerprints in the database.
Fingerprint identification has numerous practical applications. Particular fingerprints may be matched to individuals because they are distinct and unchanging. The individuality of fingerprints is based on the ridge structure and minutiae. The recognition of these landmarks — including shape, number, and location — is an automated process by which computer algorithms filter data and match a subset of individuals with a particular print. More complex analyses are then performed to identify the individual who matches the print from that subset of prospects. Overall, the accuracy of these technologies is extremely high and is considered the gold standard for individual recognition. Future technologies such as DNA fingerprinting and iris scan algorithms appear promising and may eventually replace traditional fingerprinting.
According to most professional criminal investigators, fingerprints obey three fundamental principles:
1. A fingerprint is an individual characteristic. No prints taken from different individuals have yet been found to possess identical ridge characteristics.
2. A fingerprint will remain unchanged during an individual's lifetime.
3. Fingerprints have general characteristic ridge patterns that permit them to be systematically classified.
Fingerprint analysis has been used for more than a century and remains widely used in law enforcement agencies. Because of its unique characteristics, it constitutes conclusive evidence and is a valuable tool even among advanced technologies. However, it may eventually be superseded by DNA fingerprinting, which is more sophisticated and accurate than traditional fingerprinting.
There are three types of fingerprints that may exist at crime scenes. First, visible prints are made from fingers stained with colored materials such as ink, blood, and grease. Second, plastic prints may be formed by pressing onto a soft surface such as clay, soap, or wax. Finally, a latent print is an invisible print left on an object by the body's natural greases and oils. Because latent prints cannot be seen by the naked eye, fingerprint powders, chemicals, and even lasers are used to make them visible on crime scene evidence.
In North America, one of the first successful uses of fingerprints for identification was by E. Henry in 1901, in order to stop railway workers from double collecting pay (Schulz, Reichert, Wehner, & Mattern, 2004). The Henry system derives from the pattern of ridges, which are concentrically arranged on the hands, toes, feet, and fingers. It has reliably been proven that no two individuals have identical ridge patterns, that ridge patterns are not inheritable, that ridge patterns are formed in the embryo, and that ridge patterns never change in life. After death, ridge patterns may only change as a result of decomposition. During life, they are altered only by accident, injury, burns, disease, or other unnatural causes.
The individuality of any fingerprint is based not upon the general shape or pattern it forms, but rather upon its ridge structure and specific characteristics, also known as minutiae. The recognition of these ridges, their relative number, and their approximate location on the observed print are the special characteristics that make a fingerprint a specific identifying feature of each individual. There are at least 150 individual ridge characteristics on the average fingerprint. If between 10 and 16 specific points of reference on any two corresponding fingerprints match identically, a match may be assumed. In a judicial proceeding, a point-by-point comparison must be graphically demonstrated for at least 12 different, but corresponding, points in order to prove the identity of a specific person (Maudling & Attwood, 2004).
Fingerprint classification is a technique for assigning a fingerprint into one of several pre-specified types established in the literature, thereby providing an indexing mechanism. Fingerprint classification may be viewed as a coarse-level matching of fingerprints. An input fingerprint is first matched at a coarse level to one of the pre-specified types, and then at a finer level it is compared to the subset of the database containing only that type of fingerprint. Algorithms have been developed to classify fingerprints into five classes: whorl, right loop, left loop, arch, and tented arch. The algorithm separates the number of ridges present in four directions (0°, 45°, 90°, and 135°) by filtering the central part of a fingerprint with a bank of Gabor filters (Blotta & Moler, 2004). This information is quantified to generate a FingerCode, which is used for classification. Classification is based on a two-stage classifier that uses a K-nearest neighbor classifier in the first stage and a set of neural networks in the second stage. For the five-class problem, classification accuracy of 90% is typically achieved. For the four-class problem (arch and tented arch combined into one class), classification accuracy is approximately 95%.
There are three basic fingerprint patterns: arch, loop, and whorl. More complex classification systems further subdivide these patterns into plain arches or tented arches; loops may be radial or ulnar; and whorls have additional subclassifications. However, the five most commonly used classes are whorl, right loop, left loop, arch, and tented arch. Loops make up nearly two-thirds of all fingerprints, whorls account for nearly one-third, and arches represent approximately 5–10%. These classifications are relevant in many large-scale forensic applications but are rarely used in biometric authentication.
Other features are essential to fingerprint authentication. The core is the inner point — normally in the middle of the print — around which swirls, loops, or arches center. It is frequently characterized by a ridge ending and several acutely curved ridges. Deltas are the points, normally at the lower left and right of the fingerprint, around which a triangular series of ridges centers. The ridges are also marked by pores, which appear at steady intervals. Some initial attempts have been made to use the location and distribution of pores as a means of authentication, but the resolution required to capture pores consistently is very high.
Of the two types of arches, the plain arch is the simplest of all fingerprint patterns. It is formed by ridges entering from one side of the print and exiting on the opposite side. These ridges tend to rise at the center of the pattern, forming a wavelike structure. The tented arch is similar, but instead of rising smoothly at the center, there is either a sharp upthrust or spike, or the ridges meet at an angle of less than 90 degrees. Arches do not have type lines, deltas, or cores.
Type lines are two diverging ridges that usually come into and split around an obstruction, such as a loop. A delta is the ridge point nearest the type line divergence. The core is the approximate center of the pattern. A loop must have one or more ridges entering from one side of the print, recurving, and exiting from the same side. If a loop opens toward the little finger, it is called an ulnar loop; if it opens toward the thumb, it is a radial loop. The patterned area of any loop is surrounded by two type lines, and all loops must have one delta (Saatci & Tavsanoglu, 2003).
All whorl patterns must have type lines and a minimum of two deltas. A plain whorl and a central pocket loop each have at least one ridge that makes a complete circuit, which may be in the form of a spiral, an oval, or any variant of a circular form. The main difference between these two patterns can be demonstrated by drawing an imaginary line between the two deltas: if the line touches any one of the spiral ridges, the pattern is a plain whorl; if no ridge is touched, the pattern is a central pocket loop. The double loop consists of any two loops combined into one fingerprint. Any print classified as accidental either contains two or more patterns or a pattern not covered by other categories — for example, a combination of a loop and a plain whorl, or a loop and a tented arch.
Classification of fingerprints has been used to identify individuals for over a century. This classification is a highly sophisticated task with numerous practical applications in day-to-day activities. Advanced algorithms filter data containing information about specific anatomic landmarks and match these data to fingerprints of individuals stored in databases. The accuracy of this technology is extremely high. Future technologies such as DNA fingerprinting and iris detection algorithms may ultimately supplant fingerprinting as the gold-standard modality for individual identification.
You’re 60% through this paper. Sign up to read the remaining 2 sections.
Sign Up Now — Instant Access Already a member? Log inAlways verify citation format against your institution’s current style guide requirements.