To study the rapid growth of research on organic photovoltaic (OPV) technology, development trends in the relevant research are analyzed based on CiteSpace software of text mining and visualization in scientific literature. By this analytical method, the outputs and cooperation of authors, the hot research topics, the vital references and the development trend of OPV are identified and visualized. Different from the traditional review articles by the experts on OPV, this work provides a new method of visualizing information about the development of the OPV technology research over the past decade quantitatively.

Wireless sensor networks are usually deployed in harsh and emergency scenarios, such as floods, fires, or earthquakes, where human participation to monitor and collect environmental data may be too dangerous. It can be also used for healthcare in extreme and remote environments. In such an environment, sensor nodes are faced with the risk of failure and the loss of valuable healthcare data. Therefore, fast collection and reliable storage of data becomes the two important basic topics for reliable data collection. Traditional distributed data collection protocols based on the network, such as Growth Codes proposed by Karma et al., have improved the persistence of data and the efficiency of reliable data collection in disaster scenarios. However, there are still some problems that reduce the overall efficiency. In this paper, we analyze the factors that affect the collection efficiency from a new perspective, the ratio of redundant symbols. Random feedback digestion (RFDG) model is proposed to digest the redundant symbols, similiar to our stomach digesting food, to remove redundant symbols and reduce resource consumption by using the feedback information of the already decoded code words sent by the sink node. This model can increase the valid information ratio in the network and finally increase data decoding efficiency. Three protocols are proposed in this paper according to different feedback mechanisms based on RFDG. It is shown that protocols based on RFDG outperform the growth codes protocol in data collection efficiency and reduce the delayed effect. INDEX TERMS Wireless sensor network, network coding, growth codes, data collection, feedback digestion.

No abstract

With the rapid growth of the number of images, many content-based image retrieval methods have been extensively used in our daily life. In general, image retrieval services are very expensive in terms of computing and storage. Therefore, outsourcing services to the cloud server is a good choice for image owners. However, privacy protection can become a big issue for image owners because the cloud server can only be semi-trusted. In this paper, we propose a novel image retrieval scheme. It is a ciphertext image retrieval method based on random mapping features with the bag-of-words model. After encrypting the image with Advanced Encryption Standard and block permutation, the cloud server generates random templates and then extracts the local features. All local features are clustered by k-means algorithm to form the visual word. The histogram of encrypted visual words is constructed in this way as the feature vector to represent each image. The similarity between images can be measured by the distance between feature vectors on the cloud server. Experiments and analysis prove the effect of the scheme. INDEX TERMS Image retrieval, AES encryption, BOW model, random mapping.

In the age of big data, plenty of valuable sensing data have been shared to enhance scientific innovation. However, this may cause unexpected privacy leakage. Although numerous privacy preservation techniques, such as perturbation, encryption, and anonymization, have been proposed to conceal sensitive information, it is usually at the cost of the application utility. Moreover, most of the existing works did not distinguished the underlying factors, such as data features and sampling rate, which contribute differently to utility and privacy information implied in the shared data. To well balance the application utility and privacy leakage for data sharing, we utilize mutual information and visualization techniques to analyze the impact of the underlying factors on utility and privacy, respectively, and design an interactive visualization tool to help users identify the appropriate solution to achieve the objectives of high application utility and low privacy leakage simultaneously. To illustrate the effectiveness of the proposed scheme and tool, accelerometer data collected from mobile devices have been adopted as an illustrative example. Experimental study has shown that feature selection and sampling frequency play dominant roles in reducing privacy leakage with much less reduction on utility, and the proposed visualization tool can effectively recommend the appropriate combination of features and sampling rates that can help users make decision on the trade-off between utility and privacy.

Deep neural networks are vulnerable to adversarial samples, posing potential threats to the applications deployed with deep learning models in practical conditions. A typical example is the fingerprint liveness detection module in fingerprint authentication systems. Inspired by great progress of deep learning, deep networks-based fingerprint liveness detection algorithms spring up and dominate the field. Thus, we investigate the feasibility of deceiving state-of-the-art deep networks-based fingerprint liveness detection schemes by leveraging this property in this paper. Extensive evaluations are made with three existing adversarial methods: FGSM, MI-FGSM, and Deepfool. We also proposed an adversarial attack method that enhances the robustness of adversarial fingerprint images to various transformations like rotations and flip. We demonstrate these outstanding schemes are likely to classify fake fingerprints as live fingerprints by adding tiny perturbations, even without internal details of their used model. The experimental results reveal a big loophole and threats for these schemes from a view of security, and enough attention is urgently needed to be paid on anti-adversarial not only in fingerprint liveness detection but also in all deep learning applications.

Continuous Data Protection (CDP) is an emerging scheme of data backup and recovery for Cyber Security. Compared with traditional data protection techniques, it can provide much higher reliability and allow the data state to be reverted to any point-in-time in the past. CDP continuously captures and logs every disk update, producing large amount of information, thus desperately requires an efficient mechanism of organizing such information (about the disk I/Os) in order to guarantee the performance to be acceptable. However, there are few works which address the organization of history data to improve the recovery efficiency in case of the disaster occurrences. In this paper, three methods of metadata management for Cyber Security are presented, of which two are simple implementations based on MySQL database (named DIR_MySQL and OPT_MySQL respectively) and the other one is designed with consideration of its characteristics. Experimental results show that META_CDP method is far more efficient than the other two and its performance is reasonably acceptable. Furthermore, we discuss in-depth two recovery algorithms, which are full-recovery and incremental recovery. Users can choose the optimal recovery algorithm based on the required recovery time point. INDEX TERMS CDP, data backup, recovery time, metadata management, security.

No abstract