3.1.4. Extraction of Biometric Characteristics

The result of segmentation comes in the form of a binary image (with a background value equal to 0 and the objects value is 255). An algorithm is used for searching the described characteristics of the examined object. Extraction of characteristics is essential to conserve data volume, and obtain higher processing speed. The extraction algorithm, in itself, is responsible for locating the hand contour, which is equivalent to the tested shape of the hand. Finding the contour is quite simple since the wanted line of the related point is visible from the binary image. Working of the biometric characteristics' extraction algorithm is a function of the contour found. The extraction algorithm causes the localization of the convex contour case, which is responsible for detecting the points on the finger tips. The convex contour case is a polygon, which contains the outermost points of the contour. Hence, the result is formed by a point sequence that creates the case. Consequently, the algorithm searches for the points in the values between the fingers.

The contour of the hand is again used for locating defects. Defects represent any point that is located on the contour, which is present on the part demarcated by the surrounding points. At least one defect is prevalent in between any pair of points. This follows the pattern of a convex case. Hence, the identified defects are filtered to a minimal depth (distance from the case), which eliminates a number of small defects. The only defects left are often the ones corresponding to the values between the fingers. In the next step, measurement of the biometric characteristics takes place, as displayed in Figure 3. The characteristics for each finger, and the overall characteristics of the whole hand are measured.

**Figure 3.** Depiction of all measured characteristics of an object (the human hand).
