A Preliminary Study of Tongue Prints for Biometric Authentication

The tongue is a muscular organ enclosed in the oral cavity, well protected from the external environment. It is a vital organ, which does multiple functions such as the formation of food bolus, articulation of speech and, taste perception. It is bounded by palate on the superior aspect, floor of the mouth on the inferior aspect, mandibular teeth on the lateral aspects, the pharyngeal region on the posterior aspect, and the lips in anterior aspect (1). The geometric shape of the tongue is generally constant, and the physiological surface texture does not differ much (2). The tongue is an organ that can be easily exposed for inspection and also well protected from environmental influences besides very difficult to manipulate or forge, unlike other identification systems (3).

In biometric authentication, the input sample is compared with a sample template to identify the person. Nowadays, different biometric methods like fingerprint, palm print, iris scan, signature scan, etc., are in use (4). Each of these systems has its advantages and disadvantages. The drawbacks of each system make it susceptible to a security breach. Fingerprints are the most commonly used biometric authentications, but they can change due to surgery or can be eroded and subjected to injuries and burns. A retinal scan is highly sensitive and can be affected in cases of diseases such as cataracts and astigmatism. Skin color can also be used for biometrics, but it presents with a problem as considerable differences are seen with age, burns, diseases, and use of skin creams or medications (5, 6).

Tongue print is the information carried on the exposed portion of the tongue that is the shape and texture altogether. The use of tongue print has several advantages over other biometric systems such as fingerprint, palm scan, and retinal scan (7). The tongue is unique to every person, its shape remains consistent and has invariable surface textures. The color, shape, and surface features are exclusive to every person and is further proven by the fact that there is no resemblance of tongue prints even between identical twins (8). It is not affected by external factors as it is well protected. In recent years, a lot of research is being done with the tongue as an important tool in biometric authentication. Liu et al. in 2007, were the first to propose a tongue print recognition system (9). Bade et al. also proposed tongue recognition systems based on 2D dual-tree complex wavelet (10).

A lingual impression is the impression of the dorsal surface of the tongue along with the lateral borders. This will be useful in determining the shape and the surface characteristics of the tongue. This system of identification is still in the budding stage, with very few studies done. Visual inspection and digital photography have been the time-tested methods adopted in most of the studies.

This study aimed to determine the most common tongue features, its predominant shape, and variations in males and females and different ethnic origins.

A total of two hundred and fifty Malaysian dental students participated in the study. The ethical approval for this study (IEC code: 157/2018) was obtained from the Institutional Ethics Committee of Kasturba Medical College, Manipal. The exclusion criteria were students with smoking habits and any history of systemic illness. Informed consent was obtained, followed by clinical examination of the tongue. Before the examination, the subjects were asked to rinse the mouth gently with water to remove any surface debris or food particles. Then, the subjects were asked to protrude the tongue for the observers to analyze and to take a picture of the tongue. After clinical examination, photographs (front view) were taken from a predetermined distance using a digital single length reflex (DSLR) camera. The photographs were compared in terms of morphological features such as shape, depth, and characteristics of fissures by two independent observers. The shape of the tongue was determined by using three points for reference. The reference points included the region of the tongue in contact with the commissure of the lips (when protruded outside the mouth) and the tip of the tongue, as shown in Figure 1.

Figure 1.

Reference points for determining the shape of the tongue are shown.

A total of two hundred and fifty students participated in the study. Fissures were the most common surface textural characteristic that was observed on visual inspection in the participants. It is observed that shallow fissure (69.2%) is more common than deep fissures (30.8) (Table 1). Females have a higher number of both deep (62.3%) and shallow (61.8%) fissures compared to males as seen in Table 2. In comparison to three different ethnic origins, deep fissure (35.8%) was found highly significant in the Indian race, followed by Chinese (31.3%). While the Malay race (75.7%) showed dominant shallow fissures and less deep fissures. No statistically significant difference was observed between the depth of the fissures and ethnicity, as seen in Table 3.

About the location of the fissures, fissures were not observed in 3.6% of the participants. Single pattern fissures were predominant in both males and females (69.2%) followed by two pattern fissures (26%) and three or more pattern fissures (1.2%). For single pattern fissures, central fissure, as well as vertical central fissure, are the most common fissures found in both males and females, which are seen in Table 4.

When analyzing the shape of the tongue, both U-shaped and V-shaped tongue were observed more among females than males and this observation was statistically significant (P = 0.03). Only a small amount of other shaped was observed. U-shaped was common in males (52.6%) than females (40.6%) while V-shaped was more common in females (56.1%) seen in Table 5. A significant proportion of Malay race has a U-shaped tongue (52.9%); meanwhile, V-shaped is predominantly found in Indian race (54.3%) seen in Table 6. However, no statistically significant difference was observed that ethnicity plays roles in the variation of shape of the tongue.

With fast advancing technology in recent days, there is a need to prevent impersonation, which requires a foolproof biometric tool. Biometric identification of human beings is based on the identification of physiological characteristic parameters. Tongue impression can be a promising biometric tool as the tongue is a unique organ as it varies in each person. Tongue prints cannot be forged easily. Several methods used to assess morphological features of the tongue include clinical examination, photograph, alginate impression of the tongue, computer algorithm, and sensor. In this study, clinical examination and photographs were used to identify the shape, morphology, and fissures of the tongue. Li and Wei put forward an adaptive segmentation algorithm to section tongue images (9). Various other methods also have been proposed by Stefanescu et al. (11) and Jeddy et al. (12).

In the present study, fissures were predominantly found on the dorsal surface of the tongue. Shallow fissures (69.2%) were more common than deep fissures (30.8%). Females have a higher number of both deep (62.3%) and shallow (61.8%) fissures compared to males. However, no statistically significant difference was observed between gender and depth of fissures (P = 0.942). The results stated are in agreement with those with Jeddy et al. findings (12) who stated that the difference was not statistically significant. Various patterns of fissures on the dorsum of the tongue like horizontal, multiple vertical, vertical central, central, and vertical fissures are noted. Also, the fissures can be at any location of the dorsum of the tongue and can have more than one pattern and location of an individual. In the present study, single location fissures (69.2%) were found more on of the dorsum of the tongue, and only 1.2% had fissures on multiple locations. Central fissure (22.4%) were more commonly found.

In males, U-shaped (52.6%) tongue and in females V-shaped were predominantly seen (56. 1%). This observation was statistically significant (P = 0.03). These observations were different compared with other studies. The study of Jeddy et al. (12) on Indian population indicated that in both males and females, U-shaped tongue was found predominantly, while females also had a V-shaped tongue with a sharp tip. A significant proportion of Malay race has a U-shaped tongue (52.9%); meanwhile, V-shaped is predominantly found in Indian race (54.3%). However, no statistically significant difference was observed that ethnicity plays roles in the variation of the shape of the tongue. With only a few studies on tongue prints, this study may be the pilot study regarding the influence of ethnicity on the shape of the tongue and other morphological features.