To solve the problems of traditional regression-based methods such as the loss of texture caused by face scale normalization, costing more time for retraining expanded data set to improve the generalization ability of the original model, and even potential non-convergence and incomputability, a face alignment method based on scale self-adaption and Incremental Learning (IL) was proposed. Firstly, the mapping relationship between the initial face points and the standard face points was established. Secondly, the extraction of the texture features on the original image and the normalization of the face scales were achieved via the mapping relationship. At last, incremental learning was applied to the new data set by using the existing models, which improved the generalization ability of the original model quickly. The experimental results show that the proposed method performs higher alignment accuracy than traditional regression-based methods. On the AFW (Annotated Faces in the Wild) dataset (68 feature points), the accuracy is increased by 2 to 4 percentage points; and on a 100000-level large dataset (5 feature points), the robustness is increased by 1 to 2 percentage points compared to the methods based on Deep Learning (DL). In addition, the proposed method is not only suitable for face alignment regression model, but also applicable for solving other regression models.

Since the existing 3D face acquisition technology has many restrictions on gathering scene, a 3D face reconstruction technology based on several images was proposed, and its validation was verified. First, an iterative computing model of pose and depth value estimation was proposed to implement the accurate estimation of feature depth. Then the depth values integration based on several images and shape modeling were further investigated. Finally, the Iterative Pose and Depth Optimization (IPDO) algorithm was compared with Nonlinear Least-Squares Model with Symmetry and Regularization terms (NLS1_SR) on Bosphorus database, the modeling precision was improved by 9%, and the projected image of 3D model is similar to the 2D inputted image. The experimental results show that under the condition of big pose change, the proposed recognition algorithm assisted by 3D information can improve the recognition rate of more than 50%.

Concerning the problem that existing Intra Bock Copy (IBC) algorithm is not well suitable to the screen content with a variety of different sizes and shapes of patterns, an Intra Mini-Block Copy (IMBC) algorithm was proposed to further improve coding efficiency of the screen content. Firstly, the Coding Unit (CU) was divided into a set of L mini-blocks. Secondly, each mini-block was taken as the smallest matching and replication unit, the reference mini-block which matched the mini-block best was found in the reference set R by a mini-block matching optimal selection strategy. And the location of the reference mini-block and its location in the current CU were specified by L Displacement Vectors (DVs). Finally, a prediction algorithm was firstly applied to the L DVs for eliminating the correlation between DVs before entropy encoding. Compared with the IBC algorithm, for the High Efficiency Video Coding (HEVC), Screen Content Coding (SCC) test sequences, IMBC achieved the BD-rate reduction up to 3.4%, 2.9%, 2.6% for All Intra (AI), Random Access (RA) and Low-delay B (LB) configurations in lossy coding respectively, and the Bit-rate saving up to 9.5%, 5.2%, 5.1% for AI, RA, LB configurations in lossless coding respectively. The experimental results show that IMBC algorithm can effectively improve the coding efficiency of screen image at very low additional encoding and decoding complexity.

The original inverse compositional Active Appearance Model (AAM) only does one fitting process. When the initial position is far away from the destination, the model often falls into local minimum and becomes hard to converge into the correct position. Against this problem, a dual fitting method using Multi-Resolution AAM (MR-AAM) was proposed. Firstly, the first time fitting was to locate the initial position of the face in the low-resolution AAM, and the second time fitting was to use inverse compositional algorithm in the high-resolution AAM. This method can find the exact initial position and achieve better result of facial feature point localization. The experimental results show that the proposed method performs better than traditional method in the fitting accuracy along with the real-time case.

Grid service provides a distributed computing service over heterogeneous network environment by taking the advantage of the Web Service. The relationship among grid system and network systems, especially the Web Service system, was examined in the perspective of computer network architecture. An emphasis has been given to clarification of Open Grid Service Architecture(OGSA), in which some of layers irrelevant to network functionality (outside of the scope of OSI/RM) were miss-represented in early documents.

Grid service provides a distributed computing service over heterogeneous network environment by taking the advantage of the Web Service. The relationship among grid system and network systems, especially the Web Service system, was examined in the perspective of computer network architecture. An emphasis has been given to clarification of Open Grid Service Architecture(OGSA), in which some of layers irrelevant to network functionality (outside of the scope of OSI/RM) were miss-represented in early documents.