Research On Key Techniques Of Radio Resource Allocation In Relay-enhanced Cellular Networks

Face detection and feature point positioning technology is mainly used the computer system of face image feature extraction and recognition, so as to identify the identity of the technology. Because the face is a complicated change of stereo image, in the form of a mathematical description and analysis, and because the face image is affected by the light, age, obstructions, under different conditions, therefore, the face image is put in bigger difference, thus the facial recognition is more complex than the general image recognition. Based on this the paper collected and compiled the domestic and foreign related research present situation, the research on the development of face detection, and the future trend of the development of the knowledge, and to determine the research topic and the research emphasis in this paper.Based on the study of the method of face detection and facial feature points positioning as the research object, analyzing and comparing the current practical application in the engineering project of the advantages and disadvantages of several kinds of algorithms, also has a further study of face detection and facial feature points positioning technology development process and technical index. Based on the above reasons, this article adopts the method of skin color segmentation combined with AdaBoost face positioning, to raise the detection rate and reduce the detection error rate. In this paper, the research mainly divided into six parts: the first part is introduction, mainly to a rough summary: this article on the topic research background, research purpose and meaning and the origin of the thesis, on the other hand, this topic research status at home and abroad and the development direction of the future, and to deal with the difficulty in this field; The second part is an introduction by face positioning method of skin color segmentation, which introduces the integral principle, model selection and modeling process, based on the basic knowledge of color, at the same time, aiming at several common color space for color, clustering analysis and comparing; The third part is the innovation of this paper, the process of image preprocessing. Expounds the after skin segmentation, the pretreatment process of the image to be detected, and carries on the illumination compensation processing, at the same time of structural elements in black and white images after segmentation using morphological method pretreatment and select face candidate regions; The fourth part is mainly introduces human face location algorithm based on iterative algorithm AdaBoost, and elaborates the concept, the process of the algorithm, at the same time the AdaBoost algorithm of face detection and applications such as Haar classifier in the feature selection process and the related theory of integral figure, different kinds of classifier training process and structure etc. For further information; The fifth part is the combination of skin color segmentation principle and iterative algorithm AdaBoost formed by face positioning method, combination of skin color segmentation and AdaBoost algorithm of two kinds of methods to design a new face detection system, first using skin color segmentation, and then use the AdaBoost algorithm, and choose the reason of this kind of method is this AdaBoost algorithm accuracy is higher than the accuracy of the skin color segmentation, so put it in the last test steps, so as to improve the accuracy of detection system. Then use some specific training sample, use the above selected method and the theory of training samples for testing, the test results are obtained.; The sixth part is the last part of the article, mainly on the research of the paper work summary and outlook for the future work.