Unique personal identity in a given society is very important in various aspects. Fingerprints are considered to provide the most unique feature that can differentiate one individual from the other. This paper focuses onhow fingerprints are used in determining the gender of an individual and the importance on classifying fingerprints according to gender. In today’s society, crime rates have gone high and in order to determine criminals’ proper forensic anthropology needs to be carried out. Gender classification from fingerprints helps in minimizing the time taken in identifying the criminal by narrowing the list of the suspects.
Fingerprints can be defined as marks formed by individuals’ fingers and show lines of the skin. Every individual has unique fingerprints depending on his or her genetic characteristics. The probability of two people having the same fingerprint is 1 / 1.9 X 1015 (Omidiora, Ojo, Yekini, & Tubi, 2012). This indirectly means it is virtually impossible for two people to share the same fingerprints. Fingerprints form the most dependable form of personal identification in both criminal and civil cases. Since 700 AD, they have been used in important investigations (Amayeh, Bebis, & Nicolescu, 2008).The study of the pattern of raised lines on the fingers, palm and soles is known as dermatoglyphics. Several researchers have come up with information concerning the number and density of ridges in relation to the gender of an individual. Badawi, et al. (2008) explain that the right hand has a higher ridge count as compared to the left hand. Bhardwaj, Josji, Kamra & Chowdhary (2011), researched on whether ridge density has an effect on the gender of a person. More research showed that females have a lower ridge count and breadth as compared to males (Purohit & Beg, 2011). There are three main characteristics that make fingerprints invaluable evidence in a crime scene:
Fingerprints are distinctive to every individual and in no circumstance will any two fingerprints have characteristics that are exactly the same.
Fingerprints remain same; they do not change over the course of one’s lifetime. This remains true even when the fingers are injured
Patterns formed by fingerprints can be categorized in groups that are used to narrow the variety of suspects
In the twentieth century, fingerprints were being used as lawful personal identification and even became a regular routine in forensics. Organizations for fingerprint identification were set up all over the world (Croxton, et al., 2010). Some of the fingerprint identification systems that were developed include latent fingerprint acquisition, fingerprint classification and fingerprint matching. An example is the FBI fingerprint identification that was officially set up in 1924 with a catalogue of 810,000 fingerprint cards (Federal Bureau of Investigation, 1984). The first ever fingerprinting was done by a European explorer known as Joao de Barros in the 14th century in China. To identify human fingerprints automatically so as to reduce search time and complexities involved, it is important to classify the database accurately. Accuracy obtained is normally determined by the nature of the application being used (Bhardwaj, Josji, Kamra & Chowdhary, 2011). For example, a crime related case requires higher levels of accuracy as compared to normal access control cases.
The uniqueness of fingerprints and their ability not to change form the basis of fingerprint analysis. Even though the fingerprint remains the same, when an individual grows, the growth is accounted for by the enlargement of the fingerprint patterns. In addition, fatal accidents and some diseases may change the fingerprint patterns. It is important to note that identical twins can be differentiated by fingerprint analysis. Sweat forms on the fingertips due to the presence of pores on the surfaces of the ridges of the fingers. This results in moisture being left on outside part of the object being touched by an individual thereby leaving a given print. Kaur & Mazumdar (2012) allege that the visibility of the print depends on the surface in contact. Surfaces such as metal, glass and plastic give visible prints, while paper, cardboard, and timber give invisible ones. There are two main techniques of identification of fingerprints; comparison of lifted prints and live scanning. The former, in most cases is used in forensics while the latter is mainly used in security applications (Kaur, Mazumdar, & Bhonsle, 2012).
Depending on the patterns they form, ridges on fingers can be grouped into several categories. The two main features are the ridge endings and bifurcations (Prateek & Keerthi, 2010). These features are known as minutiae and they are responsible for giving a lead for further classification and identification. In the 21st century, computer programs are used to determine whether fingerprints stored in the database match with those that are freshly produced. In order to determine the similarity, analysis is done on various levels. The fingerprint samples are first extracted with a method that is suitable; with the help of a crime scene dusting kit, a photograph or a biometric device.
The next step is to compare the algorithms to the prints. The process is subsequently done in order to identify more details up to a point where a match is found. In computer analysis, ridges, bifurcations and their positioning are being compared with the original print (Prateek & Keerthi, 2010). The work of the software and the scanners is to find as many similarities in the fingerprints as possible. It is important that characteristics to be compared are not limited in order to attain high levels of accuracy some of the features that can cause rejection of fingerprints are cracks, dirt, scars and calluses.
There are two types of fingerprint scanners; optical scanners and capacitance scanners. Optical scanners, as the name suggests use light to identify the print. This is dependent the intensity of the reflected light. Ridges are portrayed as dark while valleys as light. On the other hand, capacitance scanners use electric current to define the print (Bhardwaj, Josji, Kamra & Chowdhary, 2011). Different voltage outputs are produced by the valleys and the ridges in order to show the difference between them.
Fingerprint Gender Based Classification
Several researchers have proven that it is possible to determine the gender of an individual using fingerprint. Ridge related fingerprint parameters are the main determinant of an individual’s gender; ridge count, ridge density, ridge width, ridge thickness to valley thickness ration and fingerprint patterns. Mustanski, Bailey, & Kasper (2002) investigated on the difference in ridge count between male and female fingerprints. They took rolled fingerprints of 275 women and 275 men aged between 18 and 65 years. Their results showed that women have a higher ridge count as compared to men. Baye’s theorem states that a fingerprint having a ridge density greater than 13ridges/25 mm2 is likely to be a male whereas one with a ridge count of greater than 14 ridges/25mm2 is likely to be a female. Both Purohit & Beg (2011) converge at the point that men have a lower ridge count as compared to women.
In addition, Sangam, Krupadanam, & Anasuya (2011) conducted a study to determine bilateral asymmetry and sex difference in individuals. In their study, the main objective was note the distribution and arrangement of fingerprint patterns in both males and females. They found out that whorls were highly distributed on the thumb, the index finger and the ring finger mostly in the male species. In females, it was evident that all fingers possessed a high frequency of loops but for the ring finger. These results show a great bimanual difference between the males and the females. Whorls are less in the left hand while arches and radial loops are of a high density on the left index finger. Basically, their results evidenced there are considerable sex and bimanual differences when it comes to the distribution of fingerprint patterns.
Another key study was conducted by Mustanski, Bailey, & Kasper (2002). They came up with results that showed great difference in fingerprints from male and female individuals. Therefore, it is greatly evident that gender can be easily distinguished by fingerprints. Male and female fingerprint features can be summarized as follows:μ=13.6671, σ=4.9845, and the males, with μ=14.6914, σ=4.9336, with t-value =4.802 and ρ-value=1.685E-06 (Badawi et al., 2008)
It is important that fingerprints be analyzed in order to determine the gender of an individual (Badawi et al, 2006). Analyzing the fingerprints helps in diagnosis of inherited disorders in both prenatal and newborn babies all over the world. Examples of disorders that can be diagnosed include sickle cell anemia, thalassemia, familial Alzheimer’s, cystic fibrosis, hemophilia, and Huntington’s disease. If these diseases are detected at an early stage, they are easily treatable as compared to when they have become chronic. Moreover, when the mother is informed early enough on the possibility of her child getting some hereditary disease, she is psychologically prepared as opposed to when she gets the news unexpectedly. Furthermore, fingerprint analysis helps in developing cures for inherited disorders. Researchers use information contained DNA fingerprints to try and get cures for various inherited disorders.
Fingerprints of relatives who have a particular disorder may be used to determine the characteristic of the disorder hence getting a cure. Furthermore, fingerprints become of great aid in cases that include identifying casualties or even persons that have been reported to be missing. When the casualties are involved in the same accident, being able to identify fingerprints according to gender may prove to be of great importance. Another big importance of fingerprint identification based on gender is that it helps in forensic and criminal scenarios (Purohit & Beg, 2011). When the fingerprints are analyzed according to gender, the list of suspects is narrowed down hence reducing in investigation time. In the US, FBI and police labs have begun to use fingerprints to connect criminals with biological evidence such as blood and semen stains, hair and clothing items in contact with the suspects. Many cases in the present day are being solved using evidence from fingerprints (Amayeh, Bebis, & Nicolescu, 2008).
As soon as a person is identified as a male or a female, various biometric qualities can be used to give auxiliary data on the individual. This provides a major clue in several security and surveillance centered applications.
Proposed Method for Gender Identification
There are two domains that can be used to process an image: spatial domain and frequency domain. Both domains are based on fingerprint ridges parameters. Generally, the spatial domain involves a lot of computations while the frequency domain involves less computation and is more flexible. The proposed gender identification contains an algorithm that has various levels. First, the fingerprint image is inputted into the system in order to determine the gender. Next, various transforms are applied so as to find their fundamental frequencies. This is the analysis bit(Amayeh, Bebis, & Nicolescu, 2008). Afterwards, the threshold value is set and some identification rules are generated to differentiate male from female. Depending on the rules applied the gender is determined. In the frequency domain analysis, a fingerprint based system is created that contains an input of digital images. Various transforms are then obtained depending on the rules applied. The transforms help in determining the gender of the owners of the fingerprints. Although the threshold setting can be done manually, MATLAB program may be used.
In order to classify gender using fingerprints, ridge breadth and white lines are normally calculated using the support vector machines. Important features used in the calculation are the ratio of ridge thickness to valley thickness (RTVTR), ridge count, white lines count, concordance in pattern type between the left and right-hand fingerprints and ridge count asymmetry (Ponnarasi & Rajaram, 2012). This method is based on the perception of decision planes that help in creation and defining the decision boundaries. A given set of objects or traits are distinguished from each other using a decision plane. Support vector machine is considered to be a nonlinear classification method and yields superior grouping.
Researchers need to critically look at regional variations and genetic factors as factors that could contribute greatly to the results obtained while carrying out the experiments on gender analysis. It is possible that one’s ridge density depends on where she or he was born or maybe the family in he or she comes from. Moreover, more research on the frequency domain analysis method needs to be done in order to determine gender without having to go through a long series of actions (Sangam, Krupadanam, & Anasuya, 2011). Although fingerprint identification and classification has been widely researched it is evident that very few researchers have covered the area of fingerprint gender classification.
Undoubtedly, fingerprints are the most reliable tool for personal identification. They remain the most acceptable and trustworthy evidence in the court of law. It is impossible to find any two people possessing the same patterns of fingerprints; this applies even in the case of identical twins. Studies in relation to analyzing fingerprints according to gender have been conducted and positive results have been achieved. This is of a great advantage in many sectors, for instance, the police department is having a smooth time while dealing with fingerprint evidence as it is easy to deal with evidence that has been sorted according to gender. Criminal search time and case solving time are greatly reduced. The spatial domain and frequency domain have been proposed as two methods that can be used in determining the gender of an individual using fingerprint. Various fingerprint ridge characteristics can be used to differentiate male and female fingerprints. Moreover, when fingerprints are analyzed according to gender, there are several advantages that come with it. As discussed above, we note that matter how old one gets, fingerprint patterns do not change; they bigger the person grows the bigger the pattern becomes (Sangam, Krupadanam, Anasuya, 2011). There are a few cases that can change the pattern of an individual’s fingerprint; very long nails, fatal accidents and some particular diseases. Ridge count is high in male fingerprints as compared to female ones whereas standard deviation is high in both genders. However, there is no significant difference in the degree symmetry between the male and female fingerprints. This shows that symmetry is not a good feature to be used in determining gender in fingerprints.