With the rapid development of cloud computing, users are exposed to increasingly serious security threats such as data leakage and privacy exposure when using cloud platform services. Problems in data security, such as inaccurate screening of indicators, lack of scientific validation of reputation evaluation results are also existed. In order to solve the problems, based on cloud environment, a security reputation model using S-AlexNet convolutional neural network and dynamic game theory (SCNN-DGT) is proposed. And it is used to protect the privacy of health data in Internet of Things (IoT). Firstly, the text information of user health data is pre-classified by using S-AlexNet convolutional neural network. Then, a recommendation incentive strategy based on dynamic game theory is proposed. So that the reputation model of user health data security is built, and the evaluation system of the model is established. Finally, an experimental study is carried out to verify the validity of the model and the model index screening. It is shown by experimental results that the model can solve the problems of low reliability of health data screening index, and low accuracy of credit distinction in cloud environment. Therefore, the reliability of mobile terminals is improved, and data security and privacy protection of mobile cloud services are strengthened effectively. INDEX TERMS Health data privacy protection, cloud computing, S-AlexNet convolutional neural network, dynamic game theory, big data security reputation model, recommendation incentive strategy, Internet of Things (IoT).

The low-latency advantages of fog computing can be applied to solve high transmission latency problems of many network architectures in Internet of Vehicles. Therefore, this paper studies the application of fog computing in Internet of Vehicles. Considering that the fog network equipment deployed in Internet of Vehicles is relatively scattered, a new network architecture is proposed, which integrates cloud computing, fog computing and software defined network and other technologies. The proposed framework uses software defined network to centrally control fog network and obtains equipment performance of fog network. Furthermore, the optimal load balancing strategy is developed by communication overhead and other information. Based on time delay modeling of fog network, we study the time delay modeling of cloud-fog network and the energy consumption modeling of fog network. In addition, this paper models the selection process of data transmission network and data calculation execution server of delay-tolerant data as a partially observable Markov decision process optimization strategy in software defined Internet of Vehicles. By observing the state of system, current storage makes optimal decisions on data transmission and selection of computing nodes, thereby minimizing system overhead. Simulation results show that the proposed scheme can effectively reduce transmission delay and system overhead, improve data calculation efficiency. INDEX TERMS Delay-tolerant data transmission, software defined network, fog computing, the Internet of Vehicles, load balancing strategy, partially observable Markov decision process.

Aiming at the problems that the traditional remote sensing image classification methods cannot effectively integrate a variety of deep learning features and poor classification performance, a land resource use classification method based on a convolutional neural network (CNN) in ecological remote sensing images is proposed. In this study, a seven-layer convolution neural network is constructed, and then the two fully connected layer features of the improved CNN network training output are fused with the fifth layer pooled layer features after dimensionality reduction by principal component analysis (PCA), so as to obtain an effective remote sensing image feature of land resources based on deep learning. Further, the classification of land resources remote sensing images is completed based on a support vector machine classifier. The remote sensing images of Pingshuo mining area in Shanxi Province are used to analyze the proposed method. The results show that the edge of the recognized image is clear, the classification accuracy, misclassification rate, and kappa coefficient are 0.9472, 0.0528, and 0.9435, respectively, and the model has excellent overall performance and good classification effect.

No abstract