Biometrics is an automated system of recognizing a person based upon physical or behavior characteristics. It is a system that recognizes a person based on "who you are" instead of "what you have" or "what you know". There are four stages for security recognition:

• What you know: Password, Pin Code
• What you have: Token, Key, ID, License, Passport, Card
• What you are: Biometrics
• When and where you act: Combination of Time and Location

Things that you know or you have like password or card could be lost, stolen, or duplicated. They do not necessarily authenticate the person using the card or knowing the password, while biometrics recognize a person by his/her natural born human characteristics or his/her costumed behavior. Fingerprint is the most frequently applied physical characteristics in biometrics.

• FAR (False Acceptance Rate): Rate that a biometrics system falsely identifies different features as identical Represents security level.
• FRR (False Rejection Rate): Rate that identical features are identified as different Represents convenient level.
• STARTEK FAR : 1/100,000 ~ 1/1,000,000 STARTEK FRR: 1/30 ~ 1/100

Security level is set by the threshold matching scores. When setting security level high, i.e. setting threshold scores high, FAR reduces while FRR raises. In referring to the following picture, the cross area among X-axis, Curve B and threshold line (Area B) stands for FAR, and the cross area among X-axis, Curve A and the threshold line (Area A) stands for FRR. Then when the threshold line moves toward right, i.e. raise the security level, Area B shrinks while Area A expands. Security level will be various for different applications. The picture next to "Security Level & FAR & FRR" shows different security level demand for market segments and STARTEK's FAR vs. FRR curve.

• Smaller imaging array
• Cheaper - It requires a smaller IC chip, which significantly reduces the overall sensor cost
• Swiping motion reduces contamination buildup on the sensor, which is particularly a problem for optical sensor.

• Less accuracte - Dynamic swiping may decrease the accuracy due to increased distortion of the finger surface as it is pulled across the sensor.
• Swiping creates additional wear on the sensor surface
• More difficult to use.

Optical sensors use arrays of photodiode or phototransistor detectors to convert the energy in light incident on the detector into electrical charge. The sensor package usually includes a light-emitting-diode (LED) to illuminate the finger. 

There are two detector types used by optical sensors, charge-coupled-devices (CCD) and CMOS based optical imagers. CCD detectors are sensitive to low light levels and are capable of making excellent grayscale pictures. However, CCD fabrication is relatively expensive and neither low-light sensitivity or grayscale imaging are required for fingerprint recognition. CMOS optical imagers are manufactured in quantity and can be made with some of the image processing steps built into the chip resulting in a lower cost. 

Optical sensors for fingerprints may be affected by a number of real world factors such as stray light and surface contamination, possibly even a fingerprint impression left by a prior user. Common contaminates that deteriorate image quality include oil and dirt, scratches on the sensor surface, and condensation or ice. Some suppliers have tried to sidestep the contamination problem by directly taking a 3D image from the surface of a finger. 3D imaging technology is more hygienic but introduces a whole new set of problems and was not included in this study. 

Impostor prints are more of a problem for optical sensors than it is for other detectors because it is relatively easy to present the scanner with a convincing picture of a fingerprint. Suppliers have come up with several techniques to validate a live finger. For example optical sensors can be enhanced and made more resistant to deception with Electro-Optical imaging. This works by placing a voltage across a light-emitting polymer film. When a finger is presented, the ridges provide a ground to the polymer surface creating a small current that generating light. The fingerprint valleys remain dark so a high contrast image is produced. The polymer is directly coupled to an optical detector
 

Capacitive sensors use an array capacitor plates to image the fingerprint. Skin is conductive enough to provide a capacitive coupling with an individual capacitive element on the array. Ridges, being closer to the detector, have a higher capacitance and valleys have a lower capacitance. Some capacitive sensors apply a small voltage to the finger to enhance the signal and create better image contrast. Capacitive sensors can be sensitive to electrostatic discharge (shock) but they are insensitive to ambient lighting and are more resist contamination issues than some optical designs.