The automatic segmentation of epithelial and stromal regions in breast pathological images has very important clinical significance for the diagnosis and treatment of breast cancer. However, due to the high complexity of epithelial and stromal regions in breast tissue pathological images, it is difficult for general segmentation models to effectively train the model based on the provided segmentation labels only, and perform fast and accurate segmentation of the two regions. Therefore, based on conditional Generative Adversarial Network (cGAN), the EPithelium and Stroma segmentation conditional Generative Adversarial Network (EPScGAN) model was proposed. In EPScGAN, the discrimination mechanism of the discriminator provided a trainable loss function for the training of the generator, in order to measure the error between the segmentation result outputs of the generator and the real labels more accurately, so as to better guide the generator training. Total of 1 286 images with the size of 512×512 were randomly cropped as an experimental dataset from the expert-labeled breast pathological image datasets provided by the Netherlands Cancer Institute (NKI) and the Vancouver General Hospital (VGH). Then the dataset was divided into the training set and the test set according to the ratio of 7:3 to train and test the EPScGAN model. Experimental results show that, the mean Intersection over Union (mIoU) of the EPScGAN model on the test set is 78.12%, and compared with other 6 popular deep learning segmentation models, EPScGAN model has better segmentation performance.

Focused on the issue that the mass small file system developed based on HDFS (Hadoop Distributed File System), SMDFS (Mass Small Distributed File System), is not compatible with POSIX (Portable Operating System Interface of UNIX) constraints, a POSIX compatible technology based on local cache and an efficient metadata management technology based on temporary data cache were proposed. Firstly, the data storage area was set to realize the redirection of the file flow in the read-write mode, and then an asynchronous thread pool model was established to synchronize the data in temporary cache, thereby completing all POSIX-related file operations from the user layer to the storage layer. In addition, with the help of the metadata cache of the skip list structure, the efficiency of metadata operations such as the List directory was optimized. The test results show that, compared to the Linux client of HDFS, the performance of random read improves ten times more, the sequential read and sequential write improves about three to four times. The performance of random write can reach 20% of the local file system. Besides, the List operation efficiency of the directory improves about 10 times. However, due to the additional switching of kernel-mode and user-mode introduced by FUSE (Filesystem in Userspace), the Linux client of SMDFS3.0 has a performance penalty of about 50% compared to Java interface.

Aiming at the subjectivity and uncertainty of online service environment, as well as existing trust models cannot describe trust degree, distrust degree and uncertainty degree, simultaneously, a subjective trust model based on intuitionistic fuzzy information was proposed. Firstly, an improved approach for aggregating crisp values into Intuitionistic Fuzzy Numbers (IFN) was developed. Then, based on this approach, the direct trust IFN and the indirect trust IFN could be calculated. Furthermore, the final trust was obtained by utilizing weight distribution strategy based on intuitionstic fuzzy entropy. The experimental results demonstrate that the proposed model is effective for credit fraud, and maintains low error level when malicious entities ratio reaches 35%.

In Web service environment, the interacting entities usually cannot be predetermined and may be in different security domains. To address the access authorization for unknown users across domain borders, access control of Web service should be implemented based on domain-independent access control information but not the identities. A context-based access control policy model which can be appropriate for Web service environment was proposed. The main idea of the model was that, various access control information was abstracted and represented as a concept of context which was adopted as the center to define and perform access control policies. The context concept here acted as an intermediary between requesters and the access permissions, which was similar to the role of Role-Based Access Control (RBAC) in a way. Context-based access control policy axioms were defined based on Description Logic (DL), on the basis of these axioms, the access control policy knowledge base with the capacity of reasoning about the access control policies was put forward. Finally, the effect of access control policy enforcement was verified in Racer reasoning system, and the experiment result proved the feasibility and validity of the presented method.

Restricted Velocity Particle Swarm Optimization (RVPSO) and Self-Adaptive Velocity Particle Swarm Optimization (SAVPSO) are two recently proposed Particle Swarm Optimization (PSO) algorithms specially for solving Constrained Optimization Problem (COP), but to our knowledge, no research has been done on the applications of the two algorithms to Unconstrained Optimizations Problem (UOP). To this end, the effectiveness and performance characteristics of the two algorithms in UOP were investigated. Moreover, in view of their relatively strong conservativeness, the algorithms were improved by combining chaos factor and random strategy respectively with the search mechanism to enhance their global exploration ability. Also, the effects of different parameter settings on the performance of all these algorithms were studied. The performance of all these algorithms was evaluated on 5 typical benchmark functions. Experimental and comparison results show that the improved RVPSO is better than RVPSO in terms of robustness and global exploration ability, but it may easily get trapped into local optima when solving high-dimensional multi-modal functions; the improved SAVPSO has stronger exploration ability and faster convergence rate than improved RVPSO, and it can achieve more accurate solutions when applied to high-dimensional multi-modal functions. Therefore, the improved SAVPSO has competitive ability of global optimization, and thus is an effective algorithm for solving unconstrained optimization problems.

In order to assist in the fast and efficient diagnosis of breast cancer and provide the prognosis information for pathologists, a computer-aided diagnosis approach for automatically grading breast pathological images was proposed. In the proposed algorithm,cells of pathological images were first automatically detected by deep convolutional neural network and sliding window. Then, the algorithms of color separation based on sparse non-negative matrix factorization, marker controlled watershed, and ellipse fitting were integrated to get the boundary of each cell. A total of 203-dimensional image-derived features, including architectural features of tumor, texture and shape features of epithelial cells were extracted from the pathological images based on the detected cells and the fitted boundary. A Support Vector Machine (SVM) classifier was trained by using the extracted features to realize the automated grading of pathological images. In order to verify the proposed algorithm, a total of 49 Hematoxylin & Eosin (H&E)-stained breast pathological images obtained from 17 patients were considered. The experimental results show that,for 100 ten-fold cross-validation trials, the features with the cell shape and the spatial structure of organization of pathological image set successfully distinguish test samples of low, intermediate and high grades with classification accuracy of 90.20%. Moreover, the proposed algorithm is able to distinguish high grade, intermediate grade, and low grade patients with accuracy of 92.87%, 82.88% and 93.61%, respectively. Compared with the methods only using texture feature or architectural feature, the proposed algorithm has a higher accuracy. The proposed algorithm can accurately distinguish the grade of tumor for pathological images and the difference of accuracy between grades is small.

Aiming at the problem that the existing evaluation methods do not take into account consumers' risk attitude, a subjective trust model based on consumers' risk attitude was proposed. Firstly, the historical information of entity evaluation attributes was converted into the interval number by using set-valued statistics theory. Then, by introducing risk attitude factor, the interval evaluation matrix was transformed into the evaluation matrix with risk attitude. Subsequently, the trust level of the entity was calculated by using the idea of relative closeness. Finally, the simulation results verify that the proposed method can make better trust decisions by considering risk attitude of consumers. The simulating experiment of anti-fraud further confirms the feasibility of the subjective trust model.

Dynamic Contrast-Enhanced Magnetic Resonance Imaging (DCE-MRI) demonstrates that malignant tumors generally show faster and higher levels of enhancement than they are seen in benign or normal tissue, after an intravenous injection of the contrast agent Gd-DTPA, DCE-MRI has played important roles in diagnosis and detecting malignant tumor. However, it is still a challenge on the fast reconstruction of DCE-MR images. Based on the idea of group sparse and the theory of Compressed Sensing (CS), a conjugate gradient algorithm combined with variable density random sampling method was employed to get samples from the local k-spaces (Fourier coefficient) sampling data. Then traditional l1 norm conjugate gradient descent algorithm was extended to l2,1 norm to jointly reconstruct multiple DCE-MR images simultaneously. Compared with conventional Multi-Measurement Vector (MMV) algorithm, the proposed approach yields a faster and more accurate reconstruction result. The experimental results show that when the sampling rate is less than 40%, the joint reconstruction time based on conjugate gradient algorithm almost decreased by 30% compared with the MMV algorithm. In addition, compared with the uniform random sampling, the variable density random sampling method improves the accuracy rate about 70%.

In this paper, an abstraction of secret-sharing schemes that is accessible to a fully mechanized analysis was given. This abstraction was formalized within the applied Pi-calculus by using an equational theory that characterized the cryptographic semantics of secret share. Based on that, an encoding method from the equational theory into a convergent rewriting system was presented, which was suitable for the automated protocol verifier ProVerif. Finally, the first general soundness result for verifiable multi-secret sharing schemes was concluded: for the multi-secret sharing schemes satisfying the specified security criterion in ProVerif, the realistic adversaries modeled on multi-secret sharing schemes in Pi-calculus can simulate the ideal adversaries in verifier ProVerif, which means that realistic adversaries and ideal adversaries are indistinguishable.

Under the cloud computing environment, it has become a significant problem to decrease the power consumption while the makespan is shortened in the process of scheduling resource. Thus, this paper made span and power consumption as the optimization objectives and established power-aware resource scheduling model, then improved the Non-dominated Sorting Genetic Algorithm Ⅱ (NSGA-Ⅱ) by adopting special initialization and the learning algorithm, to solve the problem of power-aware scheduling. Consequently, the simulation results prove that the proposed scheduling algorithm not only shortens the makespan, but also decreases the power consumption effectively.

Recognition of hand-written ballot symbols is crucial in vote-counting system based on image understanding. To improve the accuracy of ballot symbol recognition, a symbol recognition approach based on run feature was proposed. Firstly, the concept of run was presented and a discriminant model was established based on run features of ballot symbols. Secondly, the relative positions of runs were described with a ternary tree. In addition, the main run regions were extracted in noisy environment by merging run regions, and an approach of recognizing ambiguous symbols was also given. Finally, the experimental results show that the run features can accurately reflect the geometrical features of ballot symbols. The proposed approach achieves high accuracy, and its accuracy is 6.07% higher than that of template matching algorithm.

This paper improved the algorithm of generating transitions and searching for path, and proposed a path editing method based on motion graphs. With regard to generating transitions, this paper detected the motion clips which can be used to blend automatically by minimizing the average frame distance between blending frames, and proposed Enhanced Dynamic Time Wrapping (EDTW) algorithm to solve this optimization problem. Concerning path search in the motion graph, this paper used the area between two curves as the target function and improved the strategy of incremental search and the strategy of branch and bound. The result shows that the proposed algorithm can edit and generate the character motions that well match the paths specified by users.