Beyond CNNs: Exploiting Further Inherent Symmetries in Medical Image Segmentation.

Automatic tumor or lesion segmentation is a crucial step in medical image analysis for computer-aided diagnosis. Although the existing methods based on convolutional neural networks (CNNs) have achieved the state-of-the-art performance, many challenges still remain in medical tumor segmentation. This is because, although the human visual system can detect symmetries in 2-D images effectively, regular CNNs can only exploit translation invariance, overlooking further inherent symmetries existing in medical images, such as rotations and reflections. To solve this problem, we propose a novel group equivariant segmentation framework by encoding those inherent symmetries for learning more precise representations. First, kernel-based equivariant operations are devised on each orientation, which allows it to effectively address the gaps of learning symmetries in existing approaches. Then, to keep segmentation networks globally equivariant, we design distinctive group layers with layer-wise symmetry constraints. Finally, based on our novel framework, extensive experiments conducted on real-world clinical data demonstrate that a group equivariant Res-UNet (called GER-UNet) outperforms its regular CNN-based counterpart and the state-of-the-art segmentation methods in the tasks of hepatic tumor segmentation, COVID-19 lung infection segmentation, and retinal vessel detection. More importantly, the newly built GER-UNet also shows potential in reducing the sample complexity and the redundancy of filters, upgrading current segmentation CNNs, and delineating organs on other medical imaging modalities.

2D medical image segmentation via learning multi-scale contextual dependencies.

Automatic medical image segmentation plays an important role as a diagnostic aid in the identification of diseases and their treatment in clinical settings. Recently proposed methods based on Convolutional Neural Networks (CNNs) have demonstrated their potential in image processing tasks, including some medical image analysis tasks. Those methods can learn various feature representations with numerous weight-shared convolutional kernels, however, the missed diagnosis rate of regions of interest (ROIs) is still high in medical image segmentation. Two crucial factors behind this shortcoming, which have been overlooked, are small ROIs from medical images and the limited context information from the existing network models. In order to reduce the missed diagnosis rate of ROIs from medical images, we propose a new segmentation framework which enhances the representative capability of small ROIs (particularly in deep layers) and explicitly learns global contextual dependencies in multi-scale feature spaces. In particular, the local features and their global dependencies from each feature space are adaptively aggregated based on both the spatial and the channel dimensions. Moreover, some visualization comparisons of the learned features from our framework further boost neural networks' interpretability. Experimental results show that, in comparison to some popular medical image segmentation and general image segmentation methods, our proposed framework achieves the state-of-the-art performance on the liver tumor segmentation task with 91.18% Sensitivity, the COVID-19 lung infection segmentation task with 75.73% Sensitivity and the retinal vessel detection task with 82.68% Sensitivity. Moreover, it is possible to integrate (parts of) the proposed framework into most of the recently proposed Fully CNN-based models, in order to improve their effectiveness in medical image segmentation tasks.

Tumor attention networks: Better feature selection, better tumor segmentation.

Compared with the traditional analysis of computed tomography scans, automatic liver tumor segmentation can supply precise tumor volumes and reduce the inter-observer variability in estimating the tumor size and the tumor burden, which could further assist physicians to make better therapeutic choices for hepatic diseases and monitoring treatment. Among current mainstream segmentation approaches, multi-layer and multi-kernel convolutional neural networks (CNNs) have attracted much attention in diverse biomedical/medical image segmentation tasks with remarkable performance. However, an arbitrary stacking of feature maps makes CNNs quite inconsistent in imitating the cognition and the visual attention of human beings for a specific visual task. To mitigate the lack of a reasonable feature selection mechanism in CNNs, we exploit a novel and effective network architecture, called Tumor Attention Networks (TA-Net), for mining adaptive features by embedding Tumor Attention layers with multi-functional modules to assist the liver tumor segmentation task. In particular, each tumor attention layer can adaptively highlight valuable tumor features and suppress unrelated ones among feature maps from 3D and 2D perspectives. Moreover, an analysis of visualization results illustrates the effectiveness of our tumor attention modules and the interpretability of CNNs for liver tumor segmentation. Furthermore, we explore different arrangements of skip connections in information fusion. A deep ablation study is also conducted to illustrate the effects of different attention strategies for hepatic tumors. The results of extensive experiments demonstrate that the proposed TA-Net increases the liver tumor segmentation performance with a lower computation cost and a small parameter overhead over the state-of-the-art methods, under various evaluation metrics on clinical benchmark data. In addition, two additional medical image datasets are used to evaluate generalization capability of TA-Net, including the comparison with general semantic segmentation methods and a non-tumor segmentation task. All the program codes have been released at https://github.com/shuchao1212/TA-Net.

CTumorGAN: a unified framework for automatic computed tomography tumor segmentation.

PURPOSE: Unlike the normal organ segmentation task, automatic tumor segmentation is a more challenging task because of the existence of similar visual characteristics between tumors and their surroundings, especially on computed tomography (CT) images with severe low contrast resolution, as well as the diversity and individual characteristics of data acquisition procedures and devices. Consequently, most of the recently proposed methods have become increasingly difficult to be applied on a different tumor dataset with good results, and moreover, some tumor segmentors usually fail to generalize beyond those datasets and modalities used in their original evaluation experiments. METHODS: In order to alleviate some of the problems with the recently proposed methods, we propose a novel unified and end-to-end adversarial learning framework for automatic segmentation of any kinds of tumors from CT scans, called CTumorGAN, consisting of a Generator network and a Discriminator network. Specifically, the Generator attempts to generate segmentation results that are close to their corresponding golden standards, while the Discriminator aims to distinguish between generated samples and real tumor ground truths. More importantly, we deliberately design different modules to take into account the well-known obstacles, e.g., severe class imbalance, small tumor localization, and the label noise problem with poor expert annotation quality, and then use these modules to guide the CTumorGAN training process by utilizing multi-level supervision more effectively. RESULTS: We conduct a comprehensive evaluation on diverse loss functions for tumor segmentation and find that mean square error is more suitable for the CT tumor segmentation task. Furthermore, extensive experiments with multiple evaluation criteria on three well-established datasets, including lung tumor, kidney tumor, and liver tumor databases, also demonstrate that our CTumorGAN achieves stable and competitive performance compared with the state-of-the-art approaches for CT tumor segmentation. CONCLUSION: In order to overcome those key challenges arising from CT datasets and solve some of the main problems existing in the current deep learning-based methods, we propose a novel unified CTumorGAN framework, which can be effectively generalized to address any kinds of tumor datasets with superior performance.

Cybersecurity and privacy issues for socially integrated mobile healthcare applications operating in a multi-cloud environment.

Social media has enabled information-sharing across massively large networks of people without spending much financial resources and time that are otherwise required in the print and electronic media. Mobile-based social media applications have overwhelmingly changed the information-sharing perspective. However, with the advent of such applications at an unprecedented scale, the privacy of the information is compromised to a larger extent if breach mitigation is not adequate. Since healthcare applications are also being developed for mobile devices so that they also benefit from the power of social media, cybersecurity privacy concerns for such sensitive applications have become critical. This article discusses the architecture of a typical mobile healthcare application, in which customized privacy levels are defined for the individuals participating in the system. It then elaborates on how the communication across a social network in a multi-cloud environment can be made more secure and private, especially for healthcare applications.

A novel fused convolutional neural network for biomedical image classification.

With the advent of biomedical imaging technology, the number of captured and stored biomedical images is rapidly increasing day by day in hospitals, imaging laboratories and biomedical institutions. Therefore, more robust biomedical image analysis technology is needed to meet the requirement of the diagnosis and classification of various kinds of diseases using biomedical images. However, the current biomedical image classification methods and general non-biomedical image classifiers cannot extract more compact biomedical image features or capture the tiny differences between similar images with different types of diseases from the same category. In this paper, we propose a novel fused convolutional neural network to develop a more accurate and highly efficient classifier for biomedical images, which combines shallow layer features and deep layer features from the proposed deep neural network architecture. In the analysis, it was observed that the shallow layers provided more detailed local features, which could distinguish different diseases in the same category, while the deep layers could convey more high-level semantic information used to classify the diseases among the various categories. A detailed comparison of our approach with traditional classification algorithms and popular deep classifiers across several public biomedical image datasets showed the superior performance of our proposed method for biomedical image classification. In addition, we also evaluated the performance of our method in modality classification of medical images using the ImageCLEFmed dataset. Graphical abstract The graphical abstract shows the fused, deep convolutional neural network architecture proposed for biomedical image classification. In the architecture, we can clearly see the feature-fusing process going from shallow layers and the deep layers.

Finger-to-Heart (F2H): Authentication for Wireless Implantable Medical Devices.

Any proposal to provide security for implantable medical devices (IMDs), such as cardiac pacemakers and defibrillators, has to achieve a trade-off between security and accessibility for doctors to gain access to an IMD, especially in an emergency scenario. In this paper, we propose a finger-to-heart (F2H) IMD authentication scheme to address this trade-off between security and accessibility. This scheme utilizes a patient's fingerprint to perform authentication for gaining access to the IMD. Doctors can gain access to the IMD and perform emergency treatment by scanning the patient's finger tip instead of asking the patient for passwords/security tokens, thereby, achieving the necessary trade-off. In the scheme, an improved minutia-cylinder-code-based fingerprint authentication algorithm is proposed for the IMD by reducing the length of each feature vector and the number of query feature vectors. Experimental results show that the improved fingerprint authentication algorithm significantly reduces both the size of messages in transmission and computational overheads in the device, and thus, can be utilized to secure the IMD. Compared to existing electrocardiogram signal-based security schemes, the F2H scheme does not require the IMD to capture or process biometric traits in every access attempt since a fingerprint template is generated and stored in the IMD beforehand. As a result, the scarce resources in the IMD are conserved, making the scheme sustainable as well as energy efficient.

A novel biomedical image indexing and retrieval system via deep preference learning.

BACKGROUND AND OBJECTIVES: The traditional biomedical image retrieval methods as well as content-based image retrieval (CBIR) methods originally designed for non-biomedical images either only consider using pixel and low-level features to describe an image or use deep features to describe images but still leave a lot of room for improving both accuracy and efficiency. In this work, we propose a new approach, which exploits deep learning technology to extract the high-level and compact features from biomedical images. The deep feature extraction process leverages multiple hidden layers to capture substantial feature structures of high-resolution images and represent them at different levels of abstraction, leading to an improved performance for indexing and retrieval of biomedical images. METHODS: We exploit the current popular and multi-layered deep neural networks, namely, stacked denoising autoencoders (SDAE) and convolutional neural networks (CNN) to represent the discriminative features of biomedical images by transferring the feature representations and parameters of pre-trained deep neural networks from another domain. Moreover, in order to index all the images for finding the similarly referenced images, we also introduce preference learning technology to train and learn a kind of a preference model for the query image, which can output the similarity ranking list of images from a biomedical image database. To the best of our knowledge, this paper introduces preference learning technology for the first time into biomedical image retrieval. RESULTS: We evaluate the performance of two powerful algorithms based on our proposed system and compare them with those of popular biomedical image indexing approaches and existing regular image retrieval methods with detailed experiments over several well-known public biomedical image databases. Based on different criteria for the evaluation of retrieval performance, experimental results demonstrate that our proposed algorithms outperform the state-of-the-art techniques in indexing biomedical images. CONCLUSIONS: We propose a novel and automated indexing system based on deep preference learning to characterize biomedical images for developing computer aided diagnosis (CAD) systems in healthcare. Our proposed system shows an outstanding indexing ability and high efficiency for biomedical image retrieval applications and it can be used to collect and annotate the high-resolution images in a biomedical database for further biomedical image research and applications.

Fast and simple high-capacity quantum cryptography with error detection.

Quantum cryptography is commonly used to generate fresh secure keys with quantum signal transmission for instant use between two parties. However, research shows that the relatively low key generation rate hinders its practical use where a symmetric cryptography component consumes the shared key. That is, the security of the symmetric cryptography demands frequent rate of key updates, which leads to a higher consumption of the internal one-time-pad communication bandwidth, since it requires the length of the key to be as long as that of the secret. In order to alleviate these issues, we develop a matrix algorithm for fast and simple high-capacity quantum cryptography. Our scheme can achieve secure private communication with fresh keys generated from Fibonacci- and Lucas- valued orbital angular momentum (OAM) states for the seed to construct recursive Fibonacci and Lucas matrices. Moreover, the proposed matrix algorithm for quantum cryptography can ultimately be simplified to matrix multiplication, which is implemented and optimized in modern computers. Most importantly, considerably information capacity can be improved effectively and efficiently by the recursive property of Fibonacci and Lucas matrices, thereby avoiding the restriction of physical conditions, such as the communication bandwidth.

A novel end-to-end classifier using domain transferred deep convolutional neural networks for biomedical images.

BACKGROUND AND OBJECTIVES: Highly accurate classification of biomedical images is an essential task in the clinical diagnosis of numerous medical diseases identified from those images. Traditional image classification methods combined with hand-crafted image feature descriptors and various classifiers are not able to effectively improve the accuracy rate and meet the high requirements of classification of biomedical images. The same also holds true for artificial neural network models directly trained with limited biomedical images used as training data or directly used as a black box to extract the deep features based on another distant dataset. In this study, we propose a highly reliable and accurate end-to-end classifier for all kinds of biomedical images via deep learning and transfer learning. METHODS: We first apply domain transferred deep convolutional neural network for building a deep model; and then develop an overall deep learning architecture based on the raw pixels of original biomedical images using supervised training. In our model, we do not need the manual design of the feature space, seek an effective feature vector classifier or segment specific detection object and image patches, which are the main technological difficulties in the adoption of traditional image classification methods. Moreover, we do not need to be concerned with whether there are large training sets of annotated biomedical images, affordable parallel computing resources featuring GPUs or long times to wait for training a perfect deep model, which are the main problems to train deep neural networks for biomedical image classification as observed in recent works. RESULTS: With the utilization of a simple data augmentation method and fast convergence speed, our algorithm can achieve the best accuracy rate and outstanding classification ability for biomedical images. We have evaluated our classifier on several well-known public biomedical datasets and compared it with several state-of-the-art approaches. CONCLUSIONS: We propose a robust automated end-to-end classifier for biomedical images based on a domain transferred deep convolutional neural network model that shows a highly reliable and accurate performance which has been confirmed on several public biomedical image datasets.

Multiple ECG Fiducial Points-Based Random Binary Sequence Generation for Securing Wireless Body Area Networks.

Generating random binary sequences (BSes) is a fundamental requirement in cryptography. A BS is a sequence of N bits, and each bit has a value of 0 or 1. For securing sensors within wireless body area networks (WBANs), electrocardiogram (ECG)-based BS generation methods have been widely investigated in which interpulse intervals (IPIs) from each heartbeat cycle are processed to produce BSes. Using these IPI-based methods to generate a 128-bit BS in real time normally takes around half a minute. In order to improve the time efficiency of such methods, this paper presents an ECG multiple fiducial-points based binary sequence generation (MFBSG) algorithm. The technique of discrete wavelet transforms is employed to detect arrival time of these fiducial points, such as P, Q, R, S, and T peaks. Time intervals between them, including RR, RQ, RS, RP, and RT intervals, are then calculated based on this arrival time, and are used as ECG features to generate random BSes with low latency. According to our analysis on real ECG data, these ECG feature values exhibit the property of randomness and, thus, can be utilized to generate random BSes. Compared with the schemes that solely rely on IPIs to generate BSes, this MFBSG algorithm uses five feature values from one heart beat cycle, and can be up to five times faster than the solely IPI-based methods. So, it achieves a design goal of low latency. According to our analysis, the complexity of the algorithm is comparable to that of fast Fourier transforms. These randomly generated ECG BSes can be used as security keys for encryption or authentication in a WBAN system.

A Survey of Authentication Schemes in Telecare Medicine Information Systems.

E-Healthcare is an emerging field that provides mobility to its users. The protected health information of the users are stored at a remote server (Telecare Medical Information System) and can be accessed by the users at anytime. Many authentication protocols have been proposed to ensure the secure authenticated access to the Telecare Medical Information System. These protocols are designed to provide certain properties such as: anonymity, untraceability, unlinkability, privacy, confidentiality, availability and integrity. They also aim to build a key exchange mechanism, which provides security against some attacks such as: identity theft, password guessing, denial of service, impersonation and insider attacks. This paper reviews these proposed authentication protocols and discusses their strengths and weaknesses in terms of ensured security and privacy properties, and computation cost. The schemes are divided in three broad categories of one-factor, two-factor and three-factor authentication schemes. Inter-category and intra-category comparison has been performed for these schemes and based on the derived results we propose future directions and recommendations that can be very helpful to the researchers who work on the design and implementation of authentication protocols.

Hybrid threshold adaptable quantum secret sharing scheme with reverse Huffman-Fibonacci-tree coding.

With prevalent attacks in communication, sharing a secret between communicating parties is an ongoing challenge. Moreover, it is important to integrate quantum solutions with classical secret sharing schemes with low computational cost for the real world use. This paper proposes a novel hybrid threshold adaptable quantum secret sharing scheme, using an m-bonacci orbital angular momentum (OAM) pump, Lagrange interpolation polynomials, and reverse Huffman-Fibonacci-tree coding. To be exact, we employ entangled states prepared by m-bonacci sequences to detect eavesdropping. Meanwhile, we encode m-bonacci sequences in Lagrange interpolation polynomials to generate the shares of a secret with reverse Huffman-Fibonacci-tree coding. The advantages of the proposed scheme is that it can detect eavesdropping without joint quantum operations, and permits secret sharing for an arbitrary but no less than threshold-value number of classical participants with much lower bandwidth. Also, in comparison with existing quantum secret sharing schemes, it still works when there are dynamic changes, such as the unavailability of some quantum channel, the arrival of new participants and the departure of participants. Finally, we provide security analysis of the new hybrid quantum secret sharing scheme and discuss its useful features for modern applications.

Cost-Effective Encryption-Based Autonomous Routing Protocol for Efficient and Secure Wireless Sensor Networks.

The deployment of intelligent remote surveillance systems depends on wireless sensor networks (WSNs) composed of various miniature resource-constrained wireless sensor nodes. The development of routing protocols for WSNs is a major challenge because of their severe resource constraints, ad hoc topology and dynamic nature. Among those proposed routing protocols, the biology-inspired self-organized secure autonomous routing protocol (BIOSARP) involves an artificial immune system (AIS) that requires a certain amount of time to build up knowledge of neighboring nodes. The AIS algorithm uses this knowledge to distinguish between self and non-self neighboring nodes. The knowledge-building phase is a critical period in the WSN lifespan and requires active security measures. This paper proposes an enhanced BIOSARP (E-BIOSARP) that incorporates a random key encryption mechanism in a cost-effective manner to provide active security measures in WSNs. A detailed description of E-BIOSARP is presented, followed by an extensive security and performance analysis to demonstrate its efficiency. A scenario with E-BIOSARP is implemented in network simulator 2 (ns-2) and is populated with malicious nodes for analysis. Furthermore, E-BIOSARP is compared with state-of-the-art secure routing protocols in terms of processing time, delivery ratio, energy consumption, and packet overhead. The findings show that the proposed mechanism can efficiently protect WSNs from selective forwarding, brute-force or exhaustive key search, spoofing, eavesdropping, replaying or altering of routing information, cloning, acknowledgment spoofing, HELLO flood attacks, and Sybil attacks.

OWL 2 learn profile: an ontology sublanguage for the learning domain.

Many experimental ontologies have been developed for the learning domain for use at different institutions. These ontologies include different OWL/OWL 2 (Web Ontology Language) constructors. However, it is not clear which OWL 2 constructors are the most appropriate ones for designing ontologies for the learning domain. It is possible that the constructors used in these learning domain ontologies match one of the three standard OWL 2 profiles (sublanguages). To investigate whether this is the case, we have analysed a corpus of 14 ontologies designed for the learning domain. We have also compared the constructors used in these ontologies with those of the OWL 2 RL profile, one of the OWL 2 standard profiles. The results of our analysis suggest that the OWL 2 constructors used in these ontologies do not exactly match the standard OWL 2 RL profile, but form a subset of that profile which we call OWL 2 Learn.

Truthful Channel Sharing for Self Coexistence of Overlapping Medical Body Area Networks.

As defined by IEEE 802.15.6 standard, channel sharing is a potential method to coordinate inter-network interference among Medical Body Area Networks (MBANs) that are close to one another. However, channel sharing opens up new vulnerabilities as selfish MBANs may manipulate their online channel requests to gain unfair advantage over others. In this paper, we address this issue by proposing a truthful online channel sharing algorithm and a companion protocol that allocates channel efficiently and truthfully by punishing MBANs for misreporting their channel request parameters such as time, duration and bid for the channel. We first present an online channel sharing scheme for unit-length channel requests and prove that it is truthful. We then generalize our model to settings with variable-length channel requests, where we propose a critical value based channel pricing and preemption scheme. A bid adjustment procedure prevents unbeneficial preemption by artificially raising the ongoing winner's bid controlled by a penalty factor λ. Our scheme can efficiently detect selfish behaviors by monitoring a trust parameter α of each MBAN and punish MBANs from cheating by suspending their requests. Our extensive simulation results show our scheme can achieve a total profit that is more than 85% of the offline optimum method in the typical MBAN settings.

Analysis of authentication protocols in agent-based systems using labeled tableaux.

The study of multiagent systems (MASs) focuses on systems in which many intelligent agents interact with each other using communication protocols. For example, an authentication protocol is used to verify and authorize agents acting on behalf of users to protect restricted data and information. After authentication, two agents should be entitled to believe that they are communicating with each other and not with intruders. For specifying and reasoning about the security properties of authentication protocols, many researchers have proposed the use of belief logics. Since authentication protocols are designed to operate in dynamic environments, it is important to model the evolution of authentication systems through time in a systematic way. We advocate the systematic combinations of logics of beliefs and time for modeling and reasoning about evolving agent beliefs in MASs. In particular, we use a temporal belief logic called TML (+) for establishing trust theories for authentication systems and also propose a labeled tableau system for this logic. To illustrate the capabilities of TML (+), we present trust theories for several well-known authentication protocols, namely, the Lowe modified wide-mouthed frog protocol, the amended Needham-Schroeder symmetric key protocol, and Kerberos. We also show how to verify certain security properties of those protocols. With the logic TML (+) and its associated modal tableaux, we are able to reason about and verify authentication systems operating in dynamic environments.