Research On Human Recognition Based On Giant And Face

The demand of intelligence video surveillance system, both in public and private places, increases rapidly within the social development. The most important function of the system is human recognition, which is also the hotspot of research topic worldwide. Since the main technology is highly related to the biometric feature, this thesis focuses on two of the major features, giant and face. They are systematically and thoroughly discussed and analyzed.Currently, there is rarely discussion about sample quality and noise occurrence during the research in giant feature. After realizing the giant recognition based on Procrutes sharp, an algorithm of giant recognition based on the compensation for periodic samples is proposed here. The qualities of giant contours are detected firstly. Then the period of the giant process is calculated by using Fourier Transform according to the contour width information. The compensation for the unsatisfactory sample is selected due to the phase, which ensures the uniform sample distribution even after data filtration. The result shows that the method gets 15% improvement of recognition accuracy.Choosing a distinctive feature and matching criterion is the key to developing a reliable face recognition system. This paper discusses the availability of one of geometric feature invariants, scale invariant feature transform (SIFT) descriptor based face recognition. The SIFT feature description of an image is typically large and slow to computer. In most cases, the difficulty of feature matching problem is aggravated when the different face expressions and image blur exist. For abovementioned issues, this thesis proposes a new method that six interest sub-regions from the face are selected to be described and later be calculated through different weights according to their distinctiveness. The square of the similarity is used to solve the problem of data deviation. The experimental results demonstrate that the method does effectively moderate the affection by face expression. It also successfully reduces the complexity and matching time of SIFT feature sets. An example of character-oriented automatic video annotation is illustrated in the end as the application.