Abstract: AnOpenCV function library which is used more widely in digital image processingto solve problems of image processing which improve the image processingeffectiveness. One of the application is contactless palm and finger detectionfor biometric figure print where the extraction of fingertips from whole handfor enhancing the security and control less complicated for passengers. Thispaper explains the proposed skin tonedetection method which is used for skin tone segmentation of hand by using thisanother application is developed for extracting fingertip .

The 1D space map isgenerated without using the past knowledge of the host image by using thisproposed method. In the HSV color space and a binary format the segmentation andpreprocessing is performed. Figure tips are detected using the convex hulltechnique Keywords:skin tone segmentation, contours 1. INTRODUCTIONThe hand gestures, robust and skin tone recognitionfrom video is one of the important challenge in computer vision.

Based uponthese there are many application such as video survilance, face and gesturerecognition, human computer interaction, for security enhancement, biometrics. Prof Dr Ing Christof Jonietz (Author)Department. of Information TechnologySRH HochschuleHeidelbergHeidelbergGermanyChristof.

[email protected] One of the important application of the handgestures is contactless palm and finger detection for biometric fingerprintverification where it capture the whole palm and extract all the figure tipswhich improves the security, minimize spoofing and evasion, criminalinvestigation and make the control less complicated for passengers. The linesof figure print is unique because it is affected by genetic and maternalenvironment so these are differ from each other.

Based on this the figure prints that are prestoredis compared with one’s fingureprint and verify the true identity. Thereare so many color spaces exist including these RGB, HSV, YUV, YIQ, YCbCr,YIQ ,XYZ, LHC, HIS. The Most important color space is RGB where the worldview in these three color matricesbecause of presence of Luminance.

The orthogonal color space YCbCr where these represents Luminance, Chromaticblue and chromatic red as these belongs to family of television transmissioncolor spaces. This color space is mainly used in video coding and compressionas it provides excellent space for separability of Luminance and chrominance.Based upon these spaces the researchers proposed so many algorithms which werenot carried out due to their drawbacks. In this proposed skin tone segmentationalgorithm, the new color space is used that contains error signals which arefrom differentiating the gray scale map and thenon red encoded grayscale version. Due to this approach the reduction of spacedimentionality from 3D RGB to 1D space advocating its unfussiness and for realtime application the rapid classifier is constructed. The flowof algorithm is as shown in the above sequence that the hand detection, hand segmentation, figure tipdetection. First algorithm one isproposed skin tone segmentation from this algorithm we will get skin tone segmentationand here contours are used to detect thehand and by the convex hull technique implemented inorder to obtain fingertipwith this finger prints can beextracted.

II.Working Methodology: Proposed Skin Tone Detection Method: Illuminationis equally spread along RGB colors in any given color image. Thus, its impactis hardly recognized here. There are many ways to saperate these illumination.

Thetransformation matrix which is used is defined as follows in eqn 1. = 0.298936021293775390, 0.587043074451121360, 0.

140209042551032500T —– (1)Where T allows for matrix multiplication.The initial color transformation is given by eqn 2. —–(2) Here denotes the 3D matrix which has RGB vectors ofhost image. Let x is equal to 1,2,….

nas n = length(R) = length(G) = length(B). x represents matrix multiplication.This reduces the RGB color representation from 3D to1D space.The vector I(x) retains the luminance by removing the hue andsaturation. By this it is regarded as a grayscale color.

After this the it triesto get another form of luminance with out considering the R vector where most ofskin color tends in the red channel. By discarding this red color it helps forcalculating the error signal. The new vector will have G or B largest elements.i.e I^(x) = argmax(G(x),B(x)) —–(3) where xequal to {1,……n}This equation is the modified way that v values computed by HSVi.

e. Hue, Saturation and value. Only difference here is it doesn’t consider thered value during calculation. The error signal is derived by subtracting the elementmatrices generated by two equations (2) and (3).

e(x) = I(x) – I^(x) —– (4) Based upon the skin probability map (SPM) which usesexplicit threshold based skin cluster which gives lower and upper boundaries ofcluster is important for the success of this proposed technique. As shown in the below fig(1) the distribution can beeasily fit into a Gaussian cure by using Expectation Maximization(EM) Which isadmitted by the projection of data.FIG(1) Toprepresents Frequency Distribution of the dataBottom represents its Gausssian curve fit Forboundaries, Let us assume m denotes mean and s denotes standard deviation and dleft and dright denotes the distance from m on left hand sideand right hand side.m-( dleft *s) appox.

0.02511m-( dright *s) appox. 0.01177 —–(5)where dleft and dright are taken 1 and 3 sigma away from mthat covers the area under the curve.

The belowequation gives the precise empirical rule set of this work—- (6) The skin tone clusters are clearly shown with 3Dprojection of three matrices I(x),I^(x),e(x) even with the including of theluminance.