Face recognition is one of the most popular research topics in the fields of imageprocessing, pattern recognition and artificial intelligence in recent years, face detection isthe prerequisite and basis of face recognition, which directly affects the effect of facerecognition, the eye is an important part of the face, it’s accurate location is also of greatsignificance for face recognition. This paper does some research on face detection and eyelocation methods based on the relevant research results at home and abroad.The face detection method based on the improved AdaBoost algorithm is proposed.First of all, Haar-like feature is used for face feature extraction, using the integral image toquick calculate the eigenvalues, and then the AdaBoost algorithm selects some criticalfeatures to construct a set of weak classifiers, yields a strong classifier using the weakclassifiers. This paper improves the weak classifier by using a new selection method ofthreshold and bias. Finally, the improved algorithm is applied into face detection, theexperimental results show that the improved face detector has faster training speed andhigh detection accuracy and robustness.An eye location method based on isophote feature is proposed. Mainly using circularsymmetry characteristics based on isophote properties to infer eye center location, first ofall, the curvedness and all of the displacement vector were calculated according to the firstorder and the second order derivative of the image, and then the center voting mechanismis established by the curvedness, the mechanism weights the most relevant displacementvector to generate the centermap region, the maximum isocenter in the centermap is theeye location result. The experimental results show that the method can achieve accurateeye location and has robustness to illumination and pose changes. |