Learn More
U-face concept was first proposed to improve the rate of face recognition. U-face is a face recognition normalization model. U-face removes background, hair, and other factors that are not conducive to face recognition; retain only the focus of the face region. In the case of necessary human face images were rotated and histogram normalized. And then(More)
A face recognition method that based on Gabor wavelet transform and fractal is proposed, Since Gabor feature is robust to illumination and expression variations and has been successfully used in face recognition area. First, the proposed method decomposes the normalized face image by convolving the face image with multi-scale and multi-orientation Gabor(More)
Presents an effective against geometric attack robustness of digital watermarking algorithm based on improved SIFT(Scale Invariant Feature Transform). The proposal achieves watermark synchronization using improved Scale-Invariant Feature Transform; Some suitable feature points are selected to form circular patches of the carrier image. The circular patches(More)
The seismic data is a kind of high dynamic range data, which is very difficult to realize high-quality visualization. In this paper, we present a new method to scale the high-dynamic-range seismic data for visualization, which can quantize the high-dynamic-range data into the low-dynamic-range data, while preserving the high frequency characteristics of the(More)
The research of H.264/AVC parallel algorithm should be on the condition of not influencing video quality and encoding rate. Through the thorough comprehension of encoding principle and bit-stream structure, this thesis designs a H.264/AVC video encoding parallel algorithm, considering four factors about encoding rate, load balance, system architecture and(More)
This paper presents a novel Chinese vehicle license plate recognition method. The system is implemented on the binary scale picture platform. The whole process is divided into two sections: preprocessing and recognition. And preprocessing comprises binarization, plate localization and character segmentation. In the part of recognition, we introduce virtual(More)
In this paper, we proposed a method which incorporated multi-scale analysis into neural nets to solve the problem that fractal coding allows fast decoding but suffers from long encoding times. This method can reduce the computational load of fractal image coding significantly though efficient classification of image improve speed of image scan. Furthermore(More)
this paper focuses on suppressing noise while enhancing features for 3-D seismic visualization. To achieve this aim, we propose a 3D tensor based anisotropic diffusion. This algorithm uses a structure tensor to robustly estimate the local orientation of the geological structures. Then, the anisotropic diffusion is steered by these tensors, which smoothes(More)