Aiming at the monosemantic problem, the vector space model with trigger pair was built to calculate the semantic similarity. Based on this, the word-clustering was carried out. For polysemantic words, the maximum entr...
详细信息
Aiming at the monosemantic problem, the vector space model with trigger pair was built to calculate the semantic similarity. Based on this, the word-clustering was carried out. For polysemantic words, the maximum entropy model with long-distance information was applied to word sense disambiguation. Because manually tagging the test sentences with semantic tag for word sense disambiguation evaluation will bring the problem of small size and uncertainty, the evaluation was put into word clustering and Chinese word segmentation system directly. The precision of Chinese word segmentation disambiguation is 92% and the word clustering can satisfy the further requirements.
The flaw of the Padhye's TCP throughput model used in this congestion control is analyzed and a novel TCP-friendly congestion control scheme for streaming based on dynamic TCP throughput model is proposed . The pr...
详细信息
The flaw of the Padhye's TCP throughput model used in this congestion control is analyzed and a novel TCP-friendly congestion control scheme for streaming based on dynamic TCP throughput model is proposed . The proposed scheme can ensure its TCP-friendliness in dynamic network condition and has prompt responsiveness to the change of network conditions. The principle of the proposed scheme is expatiated on and the equation of the dynamic TCP throughput model is given in this paper. In the end, the advantage of the scheme is validated by simulation in NS.
This paper proposes an image recognition system, called Image Guarder, to recognize adult images automatically. Image Guarder employs a hierarchical recognition framework with 3 layers, which integrates skin color det...
详细信息
This paper proposes an image recognition system, called Image Guarder, to recognize adult images automatically. Image Guarder employs a hierarchical recognition framework with 3 layers, which integrates skin color detection, skin texture validation and image classification. A novel adaptive statistical skin color model is designed to reliably detect skin pixels. Based on skin color detection, skin regions are extracted by a smooth texture validation process. From the detected skin regions, nine ad-hoc image features are extracted to represent the image content. At last, the AdaBoost algorithm is adopted to construct an integrated classifier for the final image classification. To evaluate the performance of Image Guarder, a large image dataset consisting of 59 885 benign images and 18 320 adult images is set up. The proposed system shows an exiting performance that the precision for the adult images is 88.5%, and the precision for the benign images is 92.5% respectively. On the Pentium IV 1.5 GHz PC, the average processing speed is 5.6 frames per second for the benign images, while 1.9 frames for the adult images. Image Guarder can be used to inspect or filter adult images on the personal comp.ter or in the network transmission. It also provides a technique that supports for the network security applications.
This paper builds a kinematics model of the dual-arm six degree of freedom space robot, and deducts the generalized Jacobian matrix which describes the relationship between the motion rate of manipulator end and the r...
详细信息
This paper builds a kinematics model of the dual-arm six degree of freedom space robot, and deducts the generalized Jacobian matrix which describes the relationship between the motion rate of manipulator end and the rotation rate of joints based on this model. The deducted generalized Jacobian matrix is more simple in its model and demands less work in comp.tation than those deducted by other methods. The formula of the generalized Jacobian matrix is explicit and is easy to solve. We can obtain the generalized Jacobian matrix only if we substitute the parameters of the space robot into the deducted formula of the generalized Jacobian matrix. The method of building the model fits to adopt the resolved motion rate control. This paper gives the simulation to control the dual-arm six degree of freedom space robot based on the resolved motion rate control method.
There usually are many attributes, called small-range attributes, with small number of different values in massive relations. The number of combination values of these attributes is also very few in massive relations ...
详细信息
There usually are many attributes, called small-range attributes, with small number of different values in massive relations. The number of combination values of these attributes is also very few in massive relations so that there are a lot of repeated combination values of these attributes in massive relations. It is important to remove the repeated combination values to improve the efficiency of storing and querying massive relations. A comp.ession method for removing the repeated combination values is proposed in this paper. To comp.ess a massive relation, the method partitions the relation into two small relations: one consists of the small-range attributes and the other consists of the rest attributes. The key problem is to identify the small-range attributes. The NP-hardness of this problem is proved, and two approximate algorithms are proposed to solve this problem. The comp.ession algorithms and the query processing based on the comp.essed method are also discussed. Experimental results show that the comp.ession method has high comp.ession ratio and enhances the query processing performance.
Along with the wide use in the sci.ntific and industrial areas, Service Oriented Architecture has become the most active research direction of Grid technology. Traditional technologies overemphasize the ability of com...
详细信息
Along with the wide use in the sci.ntific and industrial areas, Service Oriented Architecture has become the most active research direction of Grid technology. Traditional technologies overemphasize the ability of comp.sition but ignore the need of collaboration. This article proposes the GSML and its related tool suite. GSML defines several concepts, enlightened by π calculus, to describe the interaction between user-side comp.nents and grid resources, and then combine them into integrated one. The related tool suite includes a comp.ser and a browser. The comp.ser helps end users to organize the data flow and the control flow in a visualized way. The browser provides the runtime environment for the Grid application described by GSML. The Grid application is defined recursively, so different end users can include other's applications in their own GSML browsers to implement collaboration.
We present a novel quantization-based digital audio watermarking scheme in wavelet domain. By quantizing a host audio's wavelet coefficients (Integer Lifting Wavelet Transform) and utilizing the characteristics of...
详细信息
We present a novel quantization-based digital audio watermarking scheme in wavelet domain. By quantizing a host audio's wavelet coefficients (Integer Lifting Wavelet Transform) and utilizing the characteristics of human auditory system (HAS), the gray image is embedded using our watermarking method. Experimental results show that the proposed watermarking scheme is inaudible and robust against.various signal processing such as noising adding, lossy comp.ession, low pass filtering, re-sampling, and re-quantifying.
This paper put forward a trust management prototype based on the generalized transitive closure. In the model, we obtained the trust evaluation according to the evaluated value of trust relationship and trust experien...
详细信息
This paper put forward a trust management prototype based on the generalized transitive closure. In the model, we obtained the trust evaluation according to the evaluated value of trust relationship and trust experience, then used it to comp.te trust by combining the values recursively, and finally obtained the integral trust relationship aggregation. We experimented on the prototype in web application environment, reviewed some factors, such as malevolent recommended trust, and evaluated the preciseness of principals and noise in trust evaluation. Data from experiments indicated that this prototype could obtain the evaluated value of trust relationship objectively and precisely.
Server redundancy technology solves the problem of availability and the performance bottleneck in the traditional distributed environment, and at the same time presents a new challenge to the system management. A kind...
详细信息
Server redundancy technology solves the problem of availability and the performance bottleneck in the traditional distributed environment, and at the same time presents a new challenge to the system management. A kind of constituting principle and working model of a distributed database server system (DDSS) is presented in this paper. To overcome the weakness of dynamic scalability of redundancy resource in configuration management of the present redundancy server system, aiming at the server DDSS (every service in the system is abstracted to an object class in the model, and multi-inst.nce of the same object class redundantly serving each other is proposed), the configuration management of redundancy resource is discussed, and adaptive configuration management policy based on mobile agent technology is proposed. With the precondition of ensuring the system availability, the system performance is improved and the resource waste is reduced. In ACM, by defining award-penalty function (as for static configuration) and user request arriving rate (as for dynamic configuration) which is the gist of configuration, redundancy inst.nces is added or deleted dynamically. Finally, the algorithmic performance is analyzed and tested, and comp.red with traditional algorithm.
A new method, namely the closed-loop pipeline, is proposed for point target detection and tracking in cloudy background. The two pipelines, a continuity filter pipeline and a test pipeline, are employed in the propose...
详细信息
A new method, namely the closed-loop pipeline, is proposed for point target detection and tracking in cloudy background. The two pipelines, a continuity filter pipeline and a test pipeline, are employed in the proposed method. At first, a preprocessing step, i.e., a local contrast thresholding, is used to detect targets, then the continuous filter using three frames correlation analysis is used to effectively remove noise. The experimental results show that the proposed method can detect and track dim point targets with arbitrary trajectories accurately, and can also predict the searching windows efficiently. The method is shown to be advantageous over the traditional one in terms of searching space, comp.tational comp.exity and clutter resistance.
暂无评论