Xuzhi Wang

A more efficient triangle rasterization algorithm implemented in FPGA

3D graphics hardware acceleration has become more and more important in mobile devices. In the paper, we review some triangle rasterization algorithm previously. On that basis, we propose a novel triangle rasterization algorithm - midpoint traversal, which reduce the number of traversal points and improve the efficiency of graphic acceleration. At last, we use FPGA to implement the algorithm, drawing a conclusion: in the equality resource, compared with others triangle rasterization algorithm based edge function, the midpoint traversal can improve the efficiency.

A analysis of mobile graphics pipeline with real-time performance

With the widely development of consumer electronics, 3D graphics have been in high demand and it has also become a very important module in embedded systems. Recent years more and more attentions have been paid on 3D graphics rendering. As we know it is very difficult to achieve real-time interactive request in the complex 3D graphics rendering on a signal CPU, therefore a dedicated graphics processor is need to enhance the rendering speed. But in embedded system, source is limited and power is supplied by battery, so that the 3D graphic processing for mobile environment can not employ the same algorithm or architecture as in the pc environment. In this paper, we review a full 3D graphics pipeline and discuss optimizations of graphics architecture have been made to overcome the shortcomings in mobile environment.

Design and implementation of indoor positioning system based on iBeacon

With the rapid increase in data and multimedia services, demand for positioning has increased especially in complex indoor environment which often needs to determine the location information of the mobile terminal. There is a lack of accuracy and robustness in current indoor positioning system. This paper designs and implements an indoor positioning system based on iBeacon. We adopt Gaussian filter and Unscented Kalman filter method to robustly extract strong signals from iBeacon device. With the extracted signals, we compared them with-in database. The goal of this paper is to design and implement a mobile-based indoor location system which has the mobile applications with the Bluetooth Low Energy technology based on the iBeacon. Using a mobile terminal our system can show position results. Moreover, our system can run on both Android systems and IOS ones. Our method has better performance compared with WiFi method. The experimental results demonstrates that the error is only within 4 meters and our system can achieve accurate and robust positioning.

A novel triangle rasterization algorithm based on edge function

With the widely used of mobile devices, such as cellular phone, personal digital assistant (PDA), mobile navigator, personal multimedia player (PMP), 3D graphics are becoming key to mobile device experience. 3D graphics hardware acceleration has become more and more important in mobile devices. In the paper, we review some triangle rasterization algorithm previously. On that basis, we propose a novel triangle rasterization algorithm - midpoint traversal, which reduces the number of traversal points and improves the efficiency of graphic acceleration. At last, we use FPGA to implement the algorithm, drawing a conclusion: in the equality resource, the midpoint traversal can improve the efficiency.

A novel hexagon diamond search algorithm for Motion Estimation

Motion Estimation is very important part in video post processing. It is the most computationally intensive module. Because it is easier to be realized, block matching algorithm (BMA) and compensation techniques have been widely used in video processing. This paper proposes a novel hexagon diamond search algorithm (NHDSA), which is carried out in two main stages. Two patterns are used to locate and refine, hexagon pattern and diamond pattern. Meanwhile, through adding the threshold judgment, the NHDSA can better deal with still and semi-still blocks in successive video frames. Taking advantage of the common international standard video, such as claire, susie and so on, lots of experiments are done. It is verified that the NHDSA can produces better quality performance, less search points and calculation complexity.

Real-time pedestrian detectionl Using OpenCL

Pedestrian detection is a challenging task, due to wide variety of appearances, especially in complex real world scenes. The use of real-time pedestrian detection is of great use for a broad range of applications in multiple domains, such as surveillance and Intelligent Transportation System. In this paper we present a fast implementation of a robust pedestrian detector by using OpenCL, which is a novel open standard for heterogeneous computing. OpenCL allows for scalability to better performance and different types of hardware, with minimal changes to the implementation. To show the portable ability and performance of the new implementation of algorithm based on HOG, the algorithms are executed on the three different platforms, including CPU + NVidia GPU and CPU + AMD GPU heterogeneous system. By using a GPU as execution device, we exploit the data parallelism opportunities of the algorithm. In pedestrian detection, HOG is a very good algorithm but long running time because of its complexity. In this paper, the HOG and SVM algorithms will be optimized with OpenCL technology to achieve the goal of real-time requirements. On a single CPU + GPU machine, we reach 36 fps on the premise of algorithms portability.

Study on rasterization algorithm for graphics acceleration system

The 3D devices have been widely applied in peoples life, for example, smart phone, internet games, high-definition video, geography navigation, etc. The large scale graphics rendering depends on the computing power of the hardware greatly, the calculation of model rasterization operations need amounts of data. It has become the bottleneck of system performance. This paper proposed the method which can accelerate graphics rasterization procedure, based on the idea of tile to meet the needs of graphical applications on embedded platform. The rules and procedures of algorithm are introduced. By using of the XUP-LX110T, the experiments are carried out. It is verified that the method should been compensated the features for the lack of resources and poor computing power of embedded platforms. It can apply smaller chip area to achieve graphics acceleration with fewer resources.

Research on the motion estimation with a novel octagon cross diamond search algorithm

It is very important that motion estimation can affect the quality of the video. Because it is easier to be realized in hardware, block matching algorithm (BMA) and compensation techniques have been widely used in video processing. But it is also the most computationally intensive module, which has limited the application. This paper proposes a novel octagon cross diamond search (NOCDS) algorithm, which is carried out in two main stages. Two patterns should be used to locate and refine, octagon pattern and diamond pattern. Taking advantage of the common international standard video, such as tennis, suzie, and so on, lots of experiments are done in successive frames. Meanwhile, for the pick signal to noise ratio (PSNR) and computation, the performance of NOCDS algorithm is compared to the classical algorithm.

Dynamic password token based on SM3 algorithm

With the development of Internet technology, dynamic password token have won more and more favor in the field of identity authentication with the good security and ease of operation. Nowadays, foreigns SHA series of algorithms of dynamic password productsare widely using in our country, which have caused a potential threaten for the localization of information security. To realize the research and development of the independent technology products based state-owned cipher algorithm will make encryption products be better safeguarded. In this paper, I select the dynamic password authentication technology based on time synchronization, and a dynamic password is designed by using the state-owned SM3 algorithm Experimental result show that this method has high security and value.

Human action recognition based on Kinect and PSO-SVM by representing 3D skeletons as points in lie group

In this paper, we propose an effective method to recognize human actions. Combined with the relationship between 3D skeleton model of joint position and particle group optimization algorithm is used to optimize the support vector machine (PSO-SVM) and depth through the Kinect sensor to obtain human 3D skeleton model, each skeletal model with 20 joints and 19 joints, the relative geometry between various body parts provides a more meaningful description than their absolute locations, we explicitly model the relative 3D geometry between different body parts in our skeletal representation. Mathematically, rigid body rotations and translations in 3D space are members of the special Euclidean group SE(3), which is a matrix Lie group. Hence, we represent the relative geometry between a pair of body parts as a point in SE(3), We then perform classification using a combination of dynamic time warping, and particle swarm optimization on support vector machine (PSO-SVM), Experimental results on three action datasets: MSR-Action 3D, UT-Kinect, Florence 3D-Action, show that the proposed representation performs better than many existing skeletal representations. The proposed approach also outperforms various state-of-the-art skeleton-based human action recognition approaches.