Zhou Li-hua

Content-Based Image Retrieval Based on the Wavelet Transform and Radon Transform

The invariant using Radon transform is constructed and a new image retrieval algorithm based on the wavelet transform and Radon transform is presented. It turns to be getting multi-scale edge images by the wavelet modulus maxima after decomposing the images by the wavelet transform in the algorithm. The Radon invariant is extracted to be the shape feature of image to the multi-scale edge images and the energy of every sub frequency band is extracted to be the texture feature of image. The Gaussian model is used to normalize the different sub-characters distance to the shape feature of image and the texture feature of image separately. The shape similarity of images and the texture similarity of images are computed by the Euclidean distance separately. Finally, the weighted sums of the shape similarity of images and the texture similarity of images can be the similarity between the querying image and other images. Experiments indicate that this method is robustness to the Gaussian noises in images similarity retrieval and more effective in the image retrieval than the other algorithms discussed in the paper.

A Cache Considering Role-Based Access Control and Trust in Privilege Management Infrastructure

PMI (privilege management infrastructure) is used to perform access control to resource in an E-commerce or E-government system. With the ever-increasing need for secure transaction, the need for systems that offer a wide variety of QoS (quality-of-service) features is also growing. In order to improve the QoS of PMI system, a cache based on RBAC (Role-based Access Control) and trust is proposed. Our system is realized based on Web service. How to design the cache based on RBAC and trust in the access control model is described in detail. The algorithm to query role permission in cache and to add records in cache is dealt with. The policy to update cache is introduced also.

Exact semismooth newton SVM

The Support vector machines can be posed as quadratic program problems in a variety of ways.This paper investigates a formulation using the two-norm for the misclassification error and appending a bias norm to objective function that leads to a positive definite quadratic program only with the nonnegative constraint under a duality construction. An unconstrained convex program problem, which minimizes a differentiable convex piecewise quadratic function, is proposed as the Lagrangian dual of the quadratic program. Then an exact semismooth Newton support vector machine (ESNSVM) is obtained to solve the program speedily. Some numerical experiments demonstrate that our algorithm is very efficient comparing with the similar algorithms such as LSVM.

Improvements of the definition of image fidelity

The defect of existing fidelity definition is discussed, by using this definition, it is found that two obviously different image can be determined to be approximate or the same. According to the defect, an improved fidelity definition is presented. The difference and relation between the two definitions are discussed and the analytic expression for deriving one definition from the other is also presented