TTP: Biometric Tolerances

WinTeam automatically assigns an ID to each new record, and uses the ID field to identify each record. You may change system-assigned ID Numbers.

Use this field to enter a unique name for the record. After a new record is created, WinTeam displays this Description with the corresponding ID number in the Add/Edit list.

Fingerprint generalization is the process of scanning multiple images of the same fingerprint and then combining the features of each entry into a single fingerprint generalization, or template. The purpose of doing this is to improve the reliability of recognizing fingerprints. By scanning a single fingerprint multiple times, the algorithm is given more information to match against.

Minutiae are specific points in a finger image. There are two main types, known as ridge endings and bifurcations. Sometimes, other details, such as the points at which scars begin or terminate, are considered minutiae.

The number and locations of the minutiae vary from finger to finger in any particular person, and from person to person for any particular finger (for example, the thumb on the left hand). When a set of finger images is obtained from an individual, the number of minutiae is recorded for each finger. The precise locations of the minutiae are also recorded, in the form of numerical coordinates, for each finger. The result is a function that can be entered and stored in a computer database. A computer can rapidly compare this function with that of anyone else in the world whose finger image has been scanned.

Min Minutia Count

During generalization, each fingerprint is required to have a minimum number of minutiae points identified. If the number is lower than the minimum, the generalization is rejected.

Threshold

When generalizing fingerprint images, each entered fingerprint must have a certain number of matching minutiae. The threshold represents how closely each scan matches the other scans performed during generalization. It essentially determines if a generalization succeeds or fails based on how closely each fingerprint matches the others.

Maximum Rotation

During generalization, the maximum rotation sets the limit to the amount of rotation (or angle) a fingerprint is allowed to have with regard to the other prints entered. The lower the rotation, the more accurately a person will have to scan their fingerprint (i.e. placing their finger exactly the same on the scanner each time). The allowed values for this setting are between 0 and 180 (inclusive).

Rotation refers to the position on the scanner. For example, if the finger is placed at "six o'clock" on the scanner one time and then the next time the finger is placed at "3 o'clock" on the scanner, this represents a rotation of 90 degrees. A 180-degree rotation would mean the finger is essentially upside down on the scanner (at "12 o'clock").

Number of Prints

Enter the number of scans per finger required during the enrollment process.

Use Quality

Select this check box in order to compare or generalize the quality of the scanned image.

Quality Threshold

When comparing or generalizing fingerprint images, the quality threshold determines how good the condition of the scanned image must be.

Matching is the process of comparing a fingerprint generalization to a recently scanned fingerprint from an Employee logging in to use a device.

Min Minutia Count

The minimum number of minutiae required to make a match between two fingerprints. If the minutiae is lower in either the scanned print or the stored print, the match will be rejected.

Threshold

When comparing two fingerprints (i.e. matching) the matching threshold determines, on average, how many invalid (i.e. not from the same finger) fingerprints out of 100 will match this print. This value should generally be configured to 2 or less (ideally .01).

Maximum Rotation

When comparing two fingerprints (i.e. matching) the maximum rotation determines how much rotational (angular) variance is allowed. If the minutiae indicate that the rotation is more than the specified limit, the print will not be considered a match (even if the minutiae match when compensated for the rotation).

Rotation refers to the position on the scanner. For example, if the finger is placed at "six o'clock" on the scanner one time and then the next time the finger is placed at "3 o'clock" on the scanner, this represents a rotation of 90 degrees. A 180-degree rotation would mean the finger is essentially upside down on the scanner (at "12 o'clock").

Indicates if the Biometric Tolerance setting is active or still in use.

Used to indicate a default Biometric Tolerance when setting up a new device.

WinTeam automatically assigns an ID to each new record, and uses the ID field to identify each record. You may change system-assigned ID Numbers.

Use this field to enter a unique name for the record. After a new record is created, WinTeam displays this Description with the corresponding ID number in the Add/Edit list.

These settings apply to both generalization and matching.

Face Confidence Threshold

This value determines how similar to a human face the image must be (how confident the software is that the image is actually a face). The higher the value, the stricter the software will be in determining the "face" attribute.

Minimal IOD

IOD = Interocular Distance. This value determines the minimum distance between the eyes an image must have in order to be captured successfully. (Default: 40; before changing this value, contact a TEAM Representative for assistance.)

Maximal IOD

IOD = Interocular Distance. This value determines the maximum distance between the eyes an image can have in order to be captured successfully. Distances outside the Minimal/Maximal Interocular Distance range will be ignored and face images will not be captured. (Default: 4,000; before changing this value, contact a TEAM Representative for assistance.)

Face Quality Threshold

This value determines how good the face image must be in order to extract facial features from it. The higher the value, the better quality the face image must have in order for features to be extracted from it. (Range: 0-255; Default: 128)

Use Liveness Detection

When this option is enabled, the software will check to see if the face images are extracted from a live image stream. (This option should never be disabled.)

Liveness Threshold

This value determines how strictly the software will check the live stream for a real face. The higher the value, the more strict the software will be in determining whether the face is real. (Range: 0-100; Default: 20)

Max Roll Angle Deviation

This value determines the maximum degree of rotation a face can have in order to be properly matched. (Default: 45)

Face image generalization is the process of capturing multiple face images via a webcam live stream then combining the features of each entry into a single face image generalization, or template.

Enroll Stream Length

This value is the number of consecutive frames the stream must contain in order for enrollment to take place successfully. (Default: 10; before adjusting this value, consult a TEAM representative for assistance.)

Template Count

This value is the number of images that must be used when generalizing face images into a composite image.

Max Records Per Template

If Liveness Detection is used, this value must be 1.

Matching is the process of comparing a face image generalization to a recently captured face image from an Employee logging in to use a device.

Matching Attempts

During the authentication process, when matching (or comparing) a captured face image and a stored face image (from enrollment), this value is the number of times the software will attempt to match the captured image with a stored image in order to authenticate.

Matching Stream Length

When using Liveness Detection, a stream of consecutive face images captured from the video stream of the webcam is required. This value is the number of consecutive frames the stream must contain in order for matching to take place during authentication. (Minimum: 10; Default: 10

Matching Threshold

The False Accept Rate is the average number of times out of 100 that the software will match a face image with the wrong person. The higher this value, the lower the FAR will be; however, the number of times the software erroneously rejects a face image will increase as well.

Indicates if the Biometric Tolerance setting is active or still in use.

Used to indicate a default Biometric Tolerance when setting up a new device.