تقييم اداء المعاملات الاحصائية في نظم التعرف على الانماط

This paper in introduces an analytic study to evaluate the performance of the statistical operators of the human recognition systems which depends on the biometrics such as face and hand. The research consists two main parts. The first uses the statistical operators to extract features from the images of face, hand, and Arabic letters. In the next part, gray scale images are transformed to binary form, then, it is encoded to one-dimensional vector as a prelude to extract five statistical attributes for each image, these attributes are: mean, median, mean absolute deviation, standard deviation, Skewness. The Microsoft excel software is used to represent and plot the extracted features via built-in charts. These charts can illustrate the ability of the statistical features to separate between samples. We depends on feature vector consists of 10 feature which are: The two experiments are applied on three databases (face, hand and Arabic letters); they show that the statistical features are very promise to recognize between patterns even in the case of too similar samples.