Grain size is one of the most important parameters for metallographic microstructure analysis, which can partly determine the material performance. The measurement of grain size is based on accurate image segmentation methods, which include traditional image processing methods and emerging machine-learning-based methods. Unfortunately, traditional image processing methods can hardly segment grains correctly from metallographic images with low contrast and blurry boundaries. Moreover, the proposed machine-learning-based methods need a large dataset to train the model and can hardly deal with the segmentation challenge of complex images with fuzzy boundaries and complex structure. In this paper, an improved U-Net model is proposed to automatically accomplish image segmentation of complex metallographic images with only a small training set. The experiments on metallographic images show the significant advantage of the method, especially for the metallographic images with low contrast, a fuzzy boundary and complex structure. Compared with other deep learning methods, the improved U-Net scored higher in ACC, MIoU, Precision, and F1 indexes, among which ACC was 0.97, MIoU was 0.752, Precision was 0.98, and F1 was 0.96. The grain size was calculated based on the segmentation according to the American Society for Testing Material (ASTM) standards, producing a satisfactory result.

Big data technique is a series of novel technologies to deal with large amounts of data from various sources. Unfortunately, it is inevitable that the data from different sources conflict with each other from the aspects of format, semantics, and value. To solve the problem of conflicts, the paper proposes data consistency theory for scientific big data, including the basic concepts, properties, and quantitative evaluation method. Data consistency can be divided into different grades as complete consistency, strong consistency, weak consistency, and conditional consistency according to consistency degree and application demand. The case study is executed on material creep testing data. The analysis results show that the theory can solve the problem of conflicts in scientific big data.

With the rapid growth in a huge number of devices connecting online, Internet of Things (IoT) is rapidly growing and getting interest of researchers. IoT enabled Wireless Sensor Network (WSN) plays a significant role to collect sensing data and transmit to central repositories. Moreover, multicasting ensures efficient group communication for disseminating same query or command to all smart devices to perform mobile service computing. It is applicable in smart home, healthcare, smart cities, and smart industries for monitoring and control. To secure such sensitive information exchange, we have considered secure group communication scenario where logical trees are maintained for each group. The main problem is unnecessary rekeying when a smart device frequently joining or leaving the network. It causes computation, communication and energy overheads. To overcome the excessive rekeying problem, we have presented a Logical Tree based Secure Mobility Management Scheme (LT-SMM) using mobile service computing in IoT. It includes the group deployment phase where smart devices securely setup a group by registering with group heads for future secure information exchange. We have presented group deployment, mobile node joining and mobile node migration protocols. Moreover, we have used chaotic map based one-way hash functions to ensure message integrity. To validate our work, extensive simulations are performed using NS 2.35. TCL code is used to configure smart devices, deploy logical tree, messaging. C language is used for algorithm implementation and messaging backend coding. Results verify the supremacy of our scheme as compared to existing tree based schemes in terms of computation, communication and energy consumption.

Recently, the extraction of valid information from big data has witnessed a growing interest. Nowadays, in social networks, large parts of websites collect user profiles to provide some valuable information through personalized recommendation. Among the available recommendation algorithms, collaborative filtering (CF) is one of the most popular algorithms due to its simple framework. However, in some practices, the computational time of CF may be unsatisfactory. Meanwhile, in some cases there are noises in data, i.e., some data are invalid, it also has a great impact on algorithm performance. To speed up the time it takes to make recommendation and tackle the noise issue more effectively, we developed a novel parallel recommender system based on CF with correntropy. Instead of traditional measures used in recommendation algorithms, the correntropy was employed to compute the similarity of two items or users to achieve insensitive performance to outliers. Moreover, to reduce the computational cost, we employed the Spark framework to facilitate parallel computing. The experiments on three datasets consisting data collected from actual social networks were conducted and the experimental results showed that for social networks application, the proposed system could effectively improve the computational time and achieve satisfactory performance though invalid data existed.

Uranium is widely used in the nuclear industry and it is well known that uranium hydride, UH₃, forms when uranium is exposed to air. The associated volume change during this transition can cause the surface region to crack, compromising structural integrity. Here, hydriding regions beneath hydride surface craters are studied by secondary ion mass spectroscopy (SIMS) and X-ray photoelectron spectroscopy (XPS). Our results indicate that strain transition regions exist, which are induced by hydrogen with a certain thickness between hydride craters and the uranium bulk. The SIMS and XPS results suggest that hydrogen exists covalently with uranium and oxygen in these transition regions. A micro-scale induction period model based on the transition region and previous hydriding models is developed. Within the so-called micro-scale induction period, hydrogen diffuses and accumulates at particular sites before reaching the critical concentration required to form stoichiometric UH₃ in the transition regions.