Search

In order to overcome the security and integrity problems of the traditional university file information resource sharing model, the paper proposes a new cloud computing-based university file information resource sharing model. Supported by the cloud computing system architecture, the use of cloud computing search algorithms to fully excavate university archives information resources to realize the comprehensive sharing of university information archives, and calculate the utility parameters of university archives information resource sharing based on the mining results of cloud computing, and build a communication consumption matrix for the sharing of archives information resources in colleges and universities, and complete the construction of the sharing model of archives information resources in colleges. The comparative experiment results show that compared with the traditional sharing model, the intrusion rate and sharing transmission loss of the constructed sharing model are significantly reduced, indicating that the constructed model can effectively realize the sharing of archive information resources in universities and ensure the integrity and safety of archives.

Brain function of the drug addict will be changed after long-term use of drugs. The article designs a method for assessing the degree of addiction to physiological indicators based on the forehead close to infrared device. The experiment paradigm is designed to include three stages: the first stage is the resting state. The second stage is task stage where the picture contains the drug picture and the neutral pictures. The third stage, resting state is using the neutral pictures. Near-infrared data collection was done in three stages of this paradigm, and then classification based on the personal information of drug users and consultation with hospital experts to determine the mild, moderate and severe classification. First, a statistical analysis was conducted to find that there are indeed differences in the three types of addiction degree. Then, the Convolutional Neural Network (CNN) model was used to classify the data of 30 people, and the classification accuracy rate reached 77.04%, which verified that the distinguished method of the three types of addiction degree is completely reasonable, and the VR closed-loop sports imaging and rehabilitation training program was designed specifically to improve the self-control training of drug addicts through sports imaging and rehabilitation training.

With the ongoing development of electric technology, the complexity of the system has extremely raised, which brings challenges for the design of test strategies. Over decades, numbers of researches has dedicated for providing efficient approaches for diagnosis with high FIR and low cost. However, most of the methods only considers sequential or paralleling test strategies, which influence the estimation on the time consumption of the optimal solution. Therefore, in this paper, a novel test optimization method is proposed for combining the sequential test strategy and parallel test strategy with a new developed technique called heuristic programming, which aims at providing the mixture test strategies for the systems. Also, the decision graph is introduced to give more accurate estimation on HEF and defines the budget impact into the heuristic programming to further improve the searching efficiency. The experiments shown that comparing with state-of-art test optimization method, the proposed method is able to provide efficient test strategies in compatible time consumption with lower time cost of the strategies. Also, with the improvement of the HEF calculation, the proposed method has high efficiency on optimization process.

Glycine max (L.) Merrill is one of the most important economic and food crops in the world. Defective rate is an important indicator affecting Glycine max (L.) Merrill surface quality factors. In this study, we introduce two kinds of color- and gradient-related visual dictionary models for discrimination of Glycine max (L.) Merrill quality. The RGB color components is firstly transferred to L*a*b* color domain. The speeded-up robust features (SURF) is performed to obtain the local gradient-related characteristic representation. The KNN clustering method is individually performed to cluster the global L*a*b* color features and local gradient-related SURF features. The clustering centers can be viewed as visual words. All the visual words are collected to generate the visual dictionary. The visual clustering features are finally discriminated by the support vector machine algorithm. The result shows that the gradient-related model achieves higher discrimination accuracy than the color-related visual dictionary model.

Over decades, testing sequence optimization has been an important part for the design of testability (DFT). However, most existed method suffers from high searching complexity. Meanwhile, with the high speed of updating with electronic technology, the cost for test points and the probability of the fault state varies due to the development of the systems. Therefore, how to update the testing sequential model comes to be important to improve the efficiency on diagnosis. Since that, in this paper, we proposed a new testing sequence optimization method based on AO* and dynamic programming to improve the efficiency on generate solutions. Also, we develop the updating method for the problem under three dynamic conditions which may change the optima solution. The experiments shows that our method can achieve better performance than several state-of-art algorithms.

The technology of hyperspectral imaging technology is proposed for fast estimation of tensile strength ratios of farmland greenhouse ethylene-vinyl acetate copolymer (EVAC) film materials on the fields. Characteristic spectral curve for a EVAC film in the band range of 978 ~1676 nm is collected by using hyperspectral imaging meter. The tensile strength parameters are detected by using H25K-S powerful machine. The chemometrics method is used to reflect the internal crosslinking characteristic spectral curves and the tensile strength rates of EVAC-based instance. The approach of support vector machine (SVM) regression generates the optimal prediction performance, where the predicted square of correlation coefficient reaches 86.63%, and the predicted root mean square error declines to 1.049. The experimental results show that it is feasible to estimate the tensile strength ratios of farmland greenhouse EVAC-based film materials by using the hyperspectral technology combined with chemometrics.