Search

Due to the complexity of the anatomical structure for human organs, medical image segmentation is always a challenging computer vision task. The Convolutional Neural Network (CNN) requires a rich feature representation, which not only needs the convolutional layers from shallow to deep,but also requires the resolution from small to large. Although CNN can be used to fuse mid-level features that are employed short-cutting, this just is a simple “shallow” connection. Thus, how to obtain useful features and how to utilize these features to improve the segmentation processes are still the key issues. In this paper, Multi-features Refinement and Aggregation (MRA) makes full use of hierarchical features by using the features fusion on several levels, and reveal the importance of refinement and aggregation of features in the medical image segmentation process. The network get low-level, high-level and even mid-level features by sampling. After aggregation and re-extraction, these features are more effectively combined. Experiment results show that our method can significantly improve segmentation accuracy compared to existing feature fusion schemes. And our approach is generalized to different backbone networks with consistent accuracy gain in brain segmentation, and it sets a new state-of-the-art in the Brat-2015 benchmarks.

Over the years, Steganography and Cryptography have been complementary techniques for enforcing security of digital data. The need for the development of robust multi-layered schemes to counter the exponential grow in the power of computing devices that can compromise security is critical in the design and implementation of security systems. Therefore, we propose a new combined steganographic and cryptographic scheme using the operators of genetic algorithm (GA) such selection, crossover and mutation, and some properties of the residue number system (RNS) with an appropriate fusing technique in order to embed encrypted text within images. The proposed scheme was tested using MatLab® R2017b and a CORE™i7 processor. Simulation results show that the proposed scheme can be deployed at one level with only the stego image containing the encrypted hidden message and at another level where the stego message is further encrypted. An analysis based on standard key metrics such as visual perception and statistical methods on steganalysis and cryptanalysis show that the proposed scheme is robust, is not complex with reduced runtime and will consume less power due to the use of residue numbers when compared to similar existing schemes.

Classical Artificial Neural Networks (ANNs) though well exploited in solving classification problems, do not model perfectly the information encoding process in the human brain because ANNs encode information using rate-based coding. However, biological neurons in the brain are known to encode information using temporal coding. In order to mimic the biological method of encoding information, various Spiking Neural Network (SNN) models have been developed. However, some of these models are limited in the number of spikes and do not leverage well on some classification problems. In order to address some of the inherent challenges associated with SNN, a multi-layer learning model for a multi-spiking network is proposed in this paper. The model exploits the temporal coding of spikes and the least-squares method to derive a weight update scheme. It also employs a spike locality concept in order to determine how the synaptic weights are to be adjusted at a particular spike time so as to minimize the learning interference, and thereby, increasing the number of spikes for learning. The performance of the model is evaluated on benchmarked classification datasets. A correlation-based metric is combined with a threshold concept to measure the classification accuracy of the model. The experimental results showed that the proposed model achieved better classification accuracy than some state-of-the-art multi-layer SNN learning models.

Deep learning-based plant disease management became a cost-effective way to improved agro-productivity. Advanced train sample generation and augmentation methods enlarge train sample size and improve feature distribution but generation and augmentation introduced sample feature discrepancy due to the generation learning process and augmentation artificial bias. We proposed a generation and geometric transformation invariant feature learning method using Siamese networks with maximum mean discrepancy loss to minimize the feature distribution discrepancies coming from the generated and augmented samples. Through variational GAN and geometric transformation, we created four dataset settings to train the proposed approach. The abundant evaluation results on the PlantVillage tomato dataset demonstrated the effectiveness of the proposed approach for the ResNet50 Siamese networks in learning generation and transformation invariant features for plant disease classification.

Forest fires will cause many serious problems, which have the characteristics of strong suddenness, wide-ranging, difficult handling and extinguishing, and strong destructiveness. Though many methods were supposed to prevent and solve the problems, there are still many problems. We supposed a model containing three modules, RealLocation, BoundaryLocation and Fire Linear, to locate the fire, determine the fire line and predict the trend of the fire line separately. We test the model and compare with the original picture. The experimental results prove that the model we proposed can greatly fit the practical application.

Attribute-based encryption (ABE) is an emergent cryptographic primitive introduced in recent years. Due to its fine-grained access policy, ABE has been widely used in cloud security to build an attribute-based access control architecture. As a promising edge technology, Mobile Edge Computing (MEC), introduced for the Internet of Mobile Things (IoMT), provides the high quality services for the end-users with mobile IoT devices. In MEC, the data is stored and processed isolately and independently in fog nodes or cloudlets located at the edge of the Internet in the vicinity of the mobile users. But in most existing ABE schemes, researchers only consider there is a single data storage in the system. So this makes the existing ABE scheme not applicable for the MEC with distributed micro data storages. To solve this prominent challenge in traditional ABE, in this paper, we propose a fully decentralized outsourced ABE scheme (FDO-ABE) acted as the access control architecture for the MEC. Our FDO-ABE scheme, based on the non-interactive multi-authority ABE (NI-MA-ABE) scheme, not only supports the multi attribute authorities (AAs) in generating the multi attribute-based private keys for user, but also provides the independent and isolated ciphertext storage for the distributed cloudlets or fog nodes. What's more, to offload the computational intensive tasks from mobile devices, we design an efficient outsourced decryption (oDec) algorithm for our FDO-ABE scheme. By using our oDec algorithm, the complex operations in decryption phase can be transmitted from mobile IoT devices to the cloudlets. Summing up, our FDO-ABE scheme is more suitable as an access control model for MEC.

The 5G technology brings the substantial improvement on the quality of services (QoS), such as higher throughput, lower latency, more stable signal and more ultra-reliable data transmission, triggering a revolution for the wireless mobile network. But in a general traffic channel in the 5G-based wireless mobile network, an attacker can detect a message transmitted over a channel, or even worse, forge or tamper with the message. Building a secure channel over the two parties is a feasible solution to this uttermost data transmission security challenge in 5G-based wireless mobile network. However, how to authentication the identities of the both parties before establishing the secure channel to fully ensure the data confidentiality and integrity during the data transmission has still been a open issue. To establish a fully secure channel, in this paper, we propose a strongly secure pairing-free certificateless authenticated key agreement (PF-CL-AKA) protocol with two-way identity-based authentication before extracting the secure session key. Our protocol is provably secure in the Lippold model, which means our protocol is still secure as long as each party of the channel has at least one uncompromised partial private term. Finally, By the theoretical analysis and simulation experiments, we can observe that our scheme is practical for the real-world applications in the 5G-based wireless mobile network.

Deep learning is used to mitigate plant disease in early stages. Improving training sample size, feature distribution and overall sample quality impacted deep learning models correctness. In this work we proposed a generic plant disease recognition architecture with a two path way Encoder GAN data generation network. This network incorporated conditional sample generation and conditional reconstruction tasks that enabled the generation of unseen data feature. Using the tomato samples extracted from PlantVillage dataset, we were able to efficiently generate plausible images that boosted plant disease recognition. We tested the performance of our approach on popular CNN architectures such as ResNet50, GoogLeNet, VGG19 and AlexNet. The experiment result proved that the approach achieved good result both in terms of generation quality and plant disease classification performance compared to existing state of the art works. Therefore, our proposed architecture is a robust candidate for plant disease recognition.

In this paper, an innovative pseudonym scheme based on Direct Anonymous Attestation (DAA) and Low-Density Parity-Check (LDPC) is proposed. For the sake of improving the efficiency of generating pseudonyms, the proposed scheme uses the LDPC code to encode the credential to produce more available versions. To detect whether a vehicle’s On-Board Unit (OBU) is compromised, the proposed scheme grants the Revocation Authority (RA) to check if the compromised OBU uses a revoked pseudonym. Compared with the existing pseudonym scheme, the proposed scheme has a higher level of security and lower computation overhead.

Space-air-ground integrated network (SAGIN) is emerged as a versatile computing and traffic architecture in recent years. Though SAGIN brings many significant benefits for modern communication and computing services, there are many unprecedented challenges in SAGIN. The one critical challenge in SAGIN is the data security. In SAGIN, because the data will be stored in cleartext on cloud, the sensitive data may suffer from the illegal access by the unauthorized users even the untrusted cloud servers (CSs). Ciphertext-policy attribute-based encryption (CP-ABE), which is a type of attribute-based encryption (ABE), has been regarded as a promising solution to the critical challenge of the data security on cloud. But there are two main blemishes in traditional CP-ABE. The first one is that there is only one attribute authority (AA) in CP-ABE. If the single AA crashs down, the whole system will be shut down. The second one is that the AA cannot effectively manage the life cycle of the users’ private keys. If a user on longer has one attribute, the AA cannot revoke the user’s private key of this attribute. This means the user can still decrypt some ciphertexts using this invalid attribute. In this paper, to solve the two flaws mentioned above, we propose a multi-authority CP-ABE (MA-CP-ABE) scheme with the dynamical key revocation (DKR). Our key revocation supports both user revocation and attribute revocation. And the our revocation is time friendly. What’s more, by using our dynamically tag-based revocation algorithm, AAs can dynamically and directly re-enable or revoke the invalid attributes to users. Finally, by evaluating and implementing our scheme, we can observe that our scheme is more comprehensive and practical for cloud applications in SAGIN.

Personal health record (PHR) is a patient-centric model of health information exchange, which greatly facilitates the storage, access, and share of personal health information. In order to share the valuable resources and reduce the operational cost, the PHR service providers would like to store the PHR applications and health information data in the cloud. The private health information may be exposed to unauthorized organizations or individuals since the patient lost the physical control of their health information. Ciphertext-policy attribute-based signcryption is a promising solution to design a cloud-assisted PHR secure sharing system. It provides fine-grained access control, confidentiality, authenticity, and sender privacy of PHR data. However, a large number of pairing and modular exponentiation computations bring heavy computational overhead during designcryption process. In order to reconcile the conflict of high computational overhead and low efficiency in the designcryption process, an outsourcing scheme is proposed in this paper. In our scheme, the heavy computations are outsourced to ciphertext transformed server, only leaving a small computational overhead for the PHR user. At the same time, the extra communication overhead in our scheme is actually tolerable. Furthermore, theoretical analysis and the desired securing properties including confidentiality, unforgeability, and verifiability have been proved formally in the random oracle model. Experimental evaluation indicates that the proposed scheme is practical and feasible.

Public key encryption supporting equality test (referred to as PKE-ET) provides the capability of testing the equivalence between two messages encrypted under different public keys. Ciphertext-policy attribute-based encryption (CP-ABE) is a promising primitive to achieve versatile and secure data sharing in the cloud computing by providing flexible one-to-many encryption. In this paper, we first initialize the concept of CP-ABE with equality test (CP-ABE-ET) by combining the notions of PKE-ET and CP-ABE. Using ABE-ET primitive, the receiver can delegate a cloud server to perform an equivalence test between two messages, which are encrypted under different access policies. During the delegated equivalence test, the cloud server is unable to obtain any knowledge of the message encrypted under either access policy. We propose a concrete CP-ABE-ET scheme using bilinear pairing and Vi`ete's formulas, and give the security proof of the proposed scheme formally in the standard model. Moreover, the theoretic analysis and experimental simulation reveal that the proposed scheme is efficient and practical.

The emergence of deep learning as a commanding technique for learning heterogeneous layers of feature representations have consequently substituted traditional machine learning algorithms which are generally poor in analyzing compound sentences. Additionally, convolutional and recurrent neural networks have auspiciously yielded state-of-the-art results in sentiment classification and Natural Language Processing (NLP). In this paper, a deep sentiment representation model through the combination of multiple Convolutional Neural Networks (CNN) kernels with Long Short-Term Memory (LSTM) is proposed for sentiment classification. Our model gains word vector representation using pre-trained Global Vectors for Word Representation (GloVe) embeddings, thereafter used as input to the CNN layer which extracts higher local text representations. Finally, Bidirectional LSTM (biLSTM) generates sentiment classification of sentence representation based on context dependent features. Our combined approach of CNN and biLSTM was experimented using the Stanford Large Movie Review Dataset (IMDB) and Stanford Sentiment Treebank Dataset (SSTB) for binary classification. The evaluation achieves outstanding results in outperforming several existing approaches with 90.4% accuracy on the Stanford Sentiment Treebank dataset and 94.8% accuracy on the Stanford Large Movie Review dataset. These results are achieved with a drastic reduction of model parameters and without a pooling layer in the CNN architecture, helping to retain local and structural information in comparison to other existing deep neural network frameworks. (Abstract)

Access control plays an important role in protecting sensitive data of an organization, by employing a mechanism that grants or revokes access to the organizational resources. It is implemented with the help of different access control models either traditional or hybrid models. In this paper, we discuss DAC, MAC, RBAC and ABAC models briefly. In addition, we propose a hybrid model that is based on reverse authorization in ABAC model. The concept of negative attributes is implemented in ABAC model that directly restricts unauthorized users. Previously, several approaches to authorized users and their authority domain in ABAC model have been implemented in literature. Though, the proposed model is unique due to a different implementation of access control in the domain of Internet of Medical Things (IoMT). This work is also discussed with respect to an example scenario. In this way, the concept of negative attributes is elaborated for implementing the negative authorization in ABAC. The comparative analysis shows that this work can perform better in some scenarios by taking advantage of negative authorization.

In recent years, the most important paradigm in online display advertising is real-time bidding (RTB). It allows advertisers to buy individual ad impressions through real-time auctions, to obtain maximum revenue. However, the existing strategies usually bid an ad impression independently, ignoring the impacts of each bid on the overall revenue during the whole ad delivery period. Thus, the recent research suggests that using the reinforcement learning (RL) framework to learn the optimal bidding strategy in RTB, based on both the immediate and future rewards. In this paper, we formulate budget constrained bidding as a model-free reinforcement learning problem, where the state space is presented by the impressions' feature parameters and the auction information, while an action is to set the bidding price. Different from the prior value-based model-free work, which suffers from the convergence problem, we learn the optimal bidding strategy by employing the policy gradient model. Additionally, we design four reward functions according to different auction results and user feedback to the learned bidding strategy more in line with the optimization objectives. We evaluate the performance of the proposed bidding strategy based on a real-world dataset, and the experimental results have demonstrated the superior performance and high efficiency compared to state-of-the-art methods.

The provision of computer-mediated communication (CMC) portal services through information technologies (IT) in higher educational institutions (HEIs) in developing countries should not be an impossible task considering the growth in information systems and an upsurge of internet users. There have been numerous efforts aimed at implementing CMC portals by HEIs but just a few could be said to be successful. The aim of this research is to develop a framework that can help to provide a better understanding of how to manage the entire implementation process so as to bring the expected advantages to institutions implementing it. Results of past research had been studied to develop the conceptual framework utilizing significant theories in the field of information system implementation and institutional change. The framework addresses the adoption, implementation as well as institutionalization stages of CMC portal implementation, and a number of institutional contextual factors influencing implementation efforts within the stages.

Accurate brain tumor segmentation is the key of clinical diagnostics and treatment planning. However, a large quantity of data produced by Magnetic resonance imaging (MRI) prevents manual segmentation in a reasonable time. So, automatic approaches are required for quick and effective segmentation. Nevertheless, the spatial and structural variability among brain tumors bring a challenge to automatically segment MRI images. In this paper, an automatic end-to-end method based on Generative Adversarial Nets (GAN) is proposed for brain tumor segmentation. This method combines the generating model with the discriminant model and takes GAN instead of conditional random fields (CRF) as high-order smoothing method. The proposed method was validated in the BRATS 2015 database, it can be proven that the proposed method achieves a competitive result and the use of GAN improve the performance of networks. Furthermore, comparing with other recent CNN-based methods, the approach only takes about 10.8s to segment a patient case.

Key distribution plays a crucial role in secure device-to-device (D2D) communication. However, due to the lack of infrastructure of D2D communication and the limited storage and computing capability of wireless devices, many key distribution methods based traditional cryptography cannot be directly applied to D2D communication scenarios. In this paper, a lightweight key distribution scheme is proposed by using acceleration sensors. Specifically, a lightweight extreme points extraction and filtering algorithm is presented in the proposed scheme to extract extreme points for key generation. Moreover, a lightweight index matching algorithm is designed in the proposed scheme to further improve the bit match rate. Furthermore, an application of the proposed scheme is developed on android system and developed by JAVA. Through extensive experiments, it is demonstrated that the proposed scheme can distribute a secure key with high entropy and low computing resources and energy consumption.

Web-Based applications are becoming more increasingly technically complex and sophisticated. The very nature of their feature-rich design and their capability to collate, process, and disseminate information over the Internet or from within an intranet makes them a popular target for attack. According to Open Web Application Security Project (OWASP) Top Ten Cheat sheet-2017, SQL Injection Attack is at peak among online attacks. This can be attributed primarily to lack of awareness on software security. Developing effective SQL injection detection approaches has been a challenge in spite of extensive research in this area. In this paper, we propose a signature based SQL injection attack detection framework by integrating fingerprinting method and Pattern Matching to distinguish genuine SQL queries from malicious queries. Our framework monitors SQL queries to the database and compares them against a dataset of signatures from known SQL injection attacks. If the fingerprint method cannot determine the legitimacy of query alone, then the Aho Corasick algorithm is invoked to ascertain whether attack signatures appear in the queries. The initial experimental results of our framework indicate the approach can identify wide variety of SQL injection attacks with negligible impact on performance.

In this paper, we investigate multi-message authentication to combat adversaries with infinite computational capacity. An authentication framework over a wiretap channel (W_1, W_2) is proposed to achieve information-theoretic security with the same key. The proposed framework bridges the two research areas in physical (PHY) layer security: secure transmission and message authentication. Specifically, the sender Alice first transmits message M to the receiver Bob over (W_1, W_2) with an error correction code; then Alice employs a hash function (i.e., ε-AWU_2 hash functions) to generate a message tag S of message M using key K, and encodes S to a codeword X^n by leveraging an existing strongly secure channel coding with exponentially small (in code length n) average probability of error; finally, Alice sends X^n over (W_1, W_2) to Bob who authenticates the received messages. We develop a theorem regarding the requirements/conditions for the authentication framework to be information-theoretic secure for authenticating a polynomial number of messages. Based on this theorem, we propose and implement an efficient and feasible authentication protocol over binary symmetric wiretap channel (BSWC) by using Linear Feedback Shifting Register based (LFSR-based) hash functions and strong secure polar code. Through extensive experiments, it is demonstrated that the proposed protocol can achieve low time cost, high authentication rate, and low authentication error rate.

Medical image registration is widely used in clinical diagnosis, treatment, quality assurance and it's also the prerequisite to realize the fusion and reconstruction. This paper presents an improved medical image registration algorithm based on mutual information. First of all, B-spline gradient operator is adopted to detect edges of the reference image and the floating image and then get binarization image respectively. Next, in order to obtain the centroid coordinates, the moments of the binarization image is calculated. According to the edge detection operator, the rotation angle of thereference image and floating image are obtained. In this paper, improved mutual information (IMI) is used as a measure ofsimilarity between the reference and floating images. Theexperimental results show that the proposed algorithm has theadvantages of low computational complexity and high accuracy, and overcomes the shortcomings of the traditional mutual information easily falling into local optimum.

Be benefited from the feature of decoupling between user interaction events and inherent asynchronous communications, the publish/subscribe (PS) system has been widely applied to many modern large-scale mission critical systems. However, massive events generated by these large-scale systems signify that trying to achieve precise matching is very difficult. Most event matching services in PS system either cause the matching inefficiency, or lead to matching inaccuracy when matching vast amount of diverse subscriptions. In this paper, an efficient matching scheme with simplified Bayesian networks approach, named MESA, is proposed to filter the inconsequential event in content-based PS system. Specifically, MESA firstly filters the events by applying probabilistic skylines, then builds the Bayesian networks for users (publishers and subscribers) in PS system, and matches the event through the probability ways finally. To evaluate the performance of MESA, we carry out experiments with real-world scenario data sets. The experimental results show that the MESA scheme is efficient, robust and accurate on large-scale uncertain data sets.

Recommendation algorithm has received extensive attention in recent years. It is very difficult to directly apply traditional recommendation algorithm to offer personalized recommendation based on mobile users' browsing behaviors and ensure security of mobile users' data. In this paper, through analysis of mobile users' browsing behaviors, we obtained the reason why traditional recommendation based on browsing behaviors of mobile users lead poor performance. Based on the conclusion, we proposed a recommendation model, denoted as DPRMBDMU, which can effective offer personalized recommendation based on browsing behaviors of mobile users and, meanwhile, protect the security of mobile users' data. Finally, we proved DPRMBDMU satisfies differential privacy in theory. And we run experiments on the dataset from real world workload to evaluate efficiency of DPRMBDMU. The experiment result verifies that the efficiency of DPRMBDMU is better than the traditional recommendation algorithm based on browsing behaviors of mobile users.

Accurate tumor segmentation is an essential and crucial step for computer-aided brain tumor diagnosis and surgical planning. Subjective segmentations are widely adopted in clinical diagnosis and treating, but they are neither accurate nor reliable. An automatical and objective system for brain tumor segmentation is strongly expected. But they are still facing some challenges such as lower segmentation accuracy, demanding a priori knowledge or requiring the human intervention. In this paper, a novel and new coarse-to-fine method is proposed to segment the brain tumor. This hierarchical framework consists of preprocessing, deep learning network based classification and post-processing. The preprocessing is used to extract image patches for each MR image and obtains the gray level sequences of image patches as the input of the deep learning network. The deep learning network based classification is implemented by a stacked auto-encoder network to extract the high level abstract feature from the input, and utilizes the extracted feature to classify image patches. After mapping the classification result to a binary image, the post-processing is implemented by a morphological filter to get the final segmentation result. In order to evaluate the proposed method, the experiment was applied to segment the brain tumor for the real patient dataset. The final performance shows that the proposed brain tumor segmentation method is more accurate and efficient.

A novel method will be proposed to automatically segment the tissue of brain in magnetic resonance (MR) images. The core idea behind this method is the mixed use of Gaussian mixture model and K-means Algorithm (GKA). In this paper, the brain tissue of MR images will be segmented into White Matter (WM), Gray Matter (GM) and Cerebrospinal Fluid (CSF) by adopting the GKA method. Both the classic Gaussian Mixture Model (GMM) clustering algorithm and the classic K-means clustering algorithm have its own shortcomings when segmenting the brain tissue. In order to improve the accuracy of segment result, the GKA fusion method has been proposed to obtain the advantages of both GMM and K-means, which is based on the characteristics of brain tissue MR images. The experiments show that the novel method can achieve a better result.

With the development and benefits of cloud computing, nowadays more and more users outsource their data to third party cloud storage servers for ease of sharing and cost saving. Ciphertext-Policy Attribute-Based Encryption (CP-ABE) is a promising tool for enabling fine-grained access control over the shared data in the cloud. However, the practical application of CP-ABE in cloud data sharing system also has its own inherent challenge to regard with user revocation. To address this challenge, the paper proposes a CP-ABE scheme which supports an effective user revocation mechanism by introducing "the essential attribute" and by considering minimally trusted proxy servers, the essential attribute must be included in both ciphertext and update-key. By excluding the revoked users from update-key (which is a part of a decryption key) and by re-encrypting the only component in ciphertexts, which is associated with the essential attribute, our scheme achieves immediate and complete user revocation mechanism in CP-ABE. The proposed scheme enables a scalable and fine-grained access control for cloud data sharing system. Our scheme provides more efficiency and security level simultaneously comparing to the existing user revocable CP-ABE scheme.