The problem of occlusion occurs during multi-target tracking may result in loss of characteristics of tracking target and thus lose the tracking targets. This paper proposes a multi-target vehicle tracking algorithm based on fusion of Embedding Coupling for One-stage Instance Segmentation (EmbedMask) and Long Short-Term Memory (LSTM) model. Firstly, the obtained real-time video data is input into EmbedMask target detection model by frame for target detection. The targets are separated from background, and traditional rectangular box detection is replaced by instance segmentation. Secondly, the maximum feature data of targets is generated by the resent convolution network, which is input into the LSTM model. The continuous data of targets is obtained by calculating and estimating the motion attitude of the tracking target. Finally, the motion and detection data of targets is input into new LSTM model layer, and the fusion calculation is used to reduce the tracking loss caused by overlapping, which can ensure the accuracy of target tracking. Experimental results on standard MOT data sets show that the proposed algorithm is robust and can be used to accurately track occluded overlapping targets.
Growth towards Industry 4.0 has a significant impact on the textile manufacturing industry. In this emerging technology the business and engineering processes are interconnected. The product traceability system is playing a vital role in each sector. Unfortunately, the traditional systems in textile manufacturing industries faced a lot of challenges due to in-house process and supply chain complexity. These systems do not provide a means for reliable and rapid response to backtrack data throughout the textile process of the product. Blockchain and Internet of Things (IoT) based processes have the capacity to overcome these challenges while deploying over the traditional product traceability system. The blockchain and IoT based system provides many benefits such as communication between product process flow to improve the performance and reduce risk, improve quality, continuous involvement of worker, product fault traceability, increase supply chain visibility, and customer’s reliability and trust-ability. In this paper, we proposed Blockchain and IoT based product traceability system. This system facilitates all stakeholders like Raw Material Suppliers, Yarn Manufacturers, Customers, and Consumers to track, monitor, and quality of the product and efficient tracking of the supply chain. This proposed solution will help textile manufacturers to improve the efficiency and quality of each product. The customer’s reliability and trust-ability will boost the manufacturer due to automated data insertion through the IoT technology, decentralized, immutable, auditable, and fault tolerance blockchain technology and traceability features.
Chebyshev map is a chaotic map frequently used in design of cryptography schemes and cryptosystems based on the hardness of the Chebyshev map-based discrete logarithm (CMDL) problem. The properties of Chebyshev map have great impact on the security of these cryptosystems. It has been known that the polynomial sequences generated by Chebyshev map defined on finite fields exhibit strong periodical features which may be utilized for cryptanalysis. This paper presents the periodical properties of Chebyshev polynomial sequences. Based on the properties, an improved cryptanalysis algorithm is proposed for the CMDL problem. It turns out that a chebyshev map-based cryptosystem using Chebyshev prime number as its modulus will have better security, where the Chebyshev prime number is defined as the prime number p satisfying that ( p + 1 ) / 2 or ( p - 1 ) / 2 is also a prime number. In support of cryptanalysis, fast algorithms to calculate the value of a Chebyshev polynomial and find the minimal period of a Chebyshev polynomial sequence are proposed, too. An example is given to show the process of cryptanalysis. Computational results have shown that only a small fraction of prime numbers are valid Chebyshev prime numbers.
Arm TrustZone technology is the most widely used system-level security framework, which provides a trusted execution environment (TEE) for embedded system SoC. This paper introduces the technical principle of TrustZone in detail, explains how to extend the security features from CPU to the whole system through various security components, and briefly introduces the secure boot, the application of TrustZone in mobile devices and the variant technology based on TrustZone. At the same time, there are many attacks on TrustZone. According to the cache architecture of TrustZone, each cache line uses a NS bit to indicate whether the line belongs to secure world or normal world. The purpose is to avoid refreshing when switching the two worlds, thus reducing the performance loss. However, it supports cache lines in the two worlds to compete and evict each other, this provides an opportunity for cache attacks, and this article details this security Vulnerability.
With population growth and climate change, traditional agriculture has been unable to meet social needs well. Meanwhile, with the development of sensor hardware equipment and machine learning technology, smart agriculture has become an important transformation direction of traditional agriculture. In this study, we are committed to the analysis and processing of sensor data collected in plant factories and we model crop development as an optimization problem with respect to certain parameters, such as yield and environmental impact, which can be optimized in an automated way. More exactly, we utilize a Markov decision process to characterize crop development, then propose a deep reinforcement learning based optimization algorithm for improved crop productivity. Several simulation experiments were carried out to optimize the yield of sugar beet in plant factories. The experimental results show that this method has a good effect and makes reasonable use of the data obtained by sensors, and provides agricultural experts with a road to more sustainable agriculture.
Smart grid has been expected to provide exquisite consumption monitoring or energy trading for its equipped abundant facilities together with two-way communication, but it inevitably leads to privacy leakage of consumption data during data retransmission. Although blind signature schemes can reduce the risk of privacy leakage due to the properties of unforgeability and blindness, most of them are traceable meaning that some malicious signers can still obtain the real information of blinded signatures. In this paper, we propose an improved blind signature scheme with untraceability to further enhance privacy protection of consumption data in smart grid. We also provide security analysis and comparative performance analysis to demonstrate its feasibility.
Considering the difficulties of cross-domain data sharing and secure access tracking between multiple organizations, this paper proposes a cross-domain data sharing model based on blockchain and the features of cloud computing storage; puts forward a data sharing protocol via digital digest matching algorithm. By matching information within digital digest, this protocol can realize the restricted sharing of data while protecting data privacy. This paper also proposes a multi-level, secure storage architecture for data according to authorization, introduces a method for adding heterogeneous data to the chain, develops a data access mechanism and a privacy protection mechanism, and provides a solution for data storage and trusted sharing in untrusted environments.
The intelligent fault diagnoses and responses of power grid planning are very important parts of the future power grid. This paper proposes an intelligent fault diagnosis and method based on GIS map and IoT. The method mainly includes the stage of automatic fault diagnosis and data analysis based on IoT sensor data, and the process of panoramic display and response optimization based on GIS. The system includes response management and control module, business development monitoring module and power outage analysis module. It can be used for panoramic map display of current response to resource distribution. Besides, it also can support automatic topology tracing and fault analysis to copy with trajectory real-time update and optimal scheduling.
The goal of Detecting Credential Stuffing is to detect the Credential Stuffing attack in time and effectively. Based on OT and Cuckoo Filter, an effective credential stuffing detection protocol is proposed in this paper. The protocol involves two servers holding a collection of suspect credentials, get the result by private set intersection. By executing this protocol, the inquirer gets the same elements of both sets without knowing anything about the other elements in the other set, and the responder does not get any additional information. The validity, correctness and safety of the scheme under the semi-honesty model are verified by theoretical analysis.
Harn has introduced a (t, n) threshold secret sharing scheme recently, in which shareholders’ shares are not disclosed in the secret reconstruction phase. The benefit is that the outside adversary cannot learn the secret even if it is recovered by more than t shareholders. Moreover, Harn has further extended this scheme into a multi-secret sharing scheme so that multiple secrets can be recovered individually at different stagies. Both schemes are claimed to achieve the perfectness property using heuristic arguments. However, in this paper, we show that the above claim is false and these schemes are not perfect. In the first scheme, the coalition of $$t-1$$ shareholders can conclude that the secret is not uniformly distributed. And in the multi-secret sharing scheme, when the public parameters satisfy some special conditions, the coalition of $$t-1$$ shareholders can even use the recovered secrets to preclude some possible values for the unrecovered secrets.
Due to the variety and rich features of live content, the live cloud platform has been widely used. However, they also bring the potential threat and huge vulnerability to the users' privacy. At present, the security mechanism is simple in almost all live streaming cloud platforms. They generally use traditional security method such as Access Control List (ACL) to prevent unauthorized users from watching live streaming. In addition, they mainly consider the security of the system itself. These security measures can only guarantee the correct operation of the platform, and cannot guarantee that the privacy information of users in the platform is not leaked. Therefore, this paper proposes a privacy-preserving algorithm to protect private data of all participants in a live streaming cloud platform, including their follow relationships, live message, etc. According to semi-trusted live streaming cloud platforms, the proposed algorithm employs key exchange over elliptic curve and advanced encryption standard (AES) to preserve user privacy. A blind matching algorithm based on the encrypted tag is proposed to match the live message and intended recipients. Finally, the security of our algorithm is analyzed in depth and the results show the proposed algorithm efficiently preserves user privacy.
With the development of cloud computing and Internet of Things, they gradually merged together to form a new system called sensor-cloud. Aiming at the characteristics of Sensing-as-a-Service of sensor-cloud, this paper studies the features of sensing data, a data mining schedule is introduced into data processing and transmission during the service process in sensor-cloud. A sequential pattern mining algorithm is used to predict data sequences that appear more frequently during the service period. By providing prediction data to system user, energy consumption would be saved, and service time would be reduced. We propose a patterns mining and matching scheme that is suitable for sensor-cloud system. By generating predicted data in different tasks to match the existing sensed data with sequential patterns, the prediction model provides a scheme for task migration in sensor-cloud. Simulate result shows that our model is quiet effective in sensor-cloud. It can reduce energy consumption and improve response time under a high accuracy.
In order to improve the utilization rate of cloud computing resources within the data center, the scheduler dynamically allocates resources according to the load of each node and migrates virtual machines. Virtual machine migration is one of the effective ways to realize the dynamic allocation of resources, and virtual machine migration will cause a certain quality of service interference to the services carried on it. We analyze the impact of virtual machine migration on service quality, study the problems of virtual machine migration timing, migration objects and migration destination, targeted optimization strategies, established an evaluation model of the impact of migration mechanism on service quality. Based on this, an effective dynamic resource scheduling strategy is proposed. Experimental results show that compared with the existing online migration strategy, our model can reduce unnecessary migration by about 33% on average while reducing migration costs by 30%. In addition, our proposed resource scheduling strategy solves the problem of insufficient resources during the subsequent migration of a heavily loaded virtual machine.
