Biblio

Missing values are an unavoidable problem for classification tasks of machine learning in medical data. With the rapid development of the medical system, large scale medical data is increasing. Missing values increase the difficulty of mining hidden but useful information in these medical datasets. Deletion and imputation methods are the most popular methods for dealing with missing values. Existing studies ignored to compare and discuss the deletion and imputation methods of missing values under the row missing rate and the total missing rate. Meanwhile, they rarely used experiment data sets that are mixed-type and large scale. In this work, medical data sets of various sizes and mixed-type are used. At the same time, performance differences of deletion and imputation methods are compared under the MCAR (Missing Completely At Random) mechanism in the baseline task using LR (Linear Regression) and SVM (Support Vector Machine) classifiers for classification with the same row and total missing rates. Experimental results show that under the MCAR missing mechanism, the performance of two types of processing methods is related to the size of datasets and missing rates. As the increasing of missing rate, the performance of two types for processing missing values decreases, but the deletion method decreases faster, and the imputation methods based on machine learning have more stable and better classification performance on average. In addition, small data sets are easily affected by processing methods of missing values.

Cloud computing has become an integral part of medical big data. The cloud has the capability to store the large data volumes has attracted more attention. The integrity and privacy of patient data are some of the issues that cloud-based medical big data should be addressed. This research work introduces data integrity auditing scheme for cloud-based medical big data. This will help minimize the risk of unauthorized access to the data. Multiple copies of the data are stored to ensure that it can be recovered quickly in case of damage. This scheme can also be used to enable doctors to easily track the changes in patients' conditions through a data block. The simulation results proved the effectiveness of the proposed scheme.

Easy Data Augmentation is originally developed for text classification tasks. It consists of four basic methods: Synonym Replacement, Random Insertion, Random Deletion, and Random Swap. They yield accuracy improvements on several deep neural network models. In this study we apply these methods to a new domain. We augment Named Entity Recognition datasets from medical domain. Although the augmentation task is much more difficult due to the nature of named entities which consist of word or word groups in the sentences, we show that we can improve the named entity recognition performance.

The objective of the work is to provide security to the medical images by embedding medical data (EPR-Electronic Patient Record) along with the image to reduce the bandwidth during communication. Reversible watermarking and Digital Signature itself will provide high security. This application mainly used in tele-surgery (Medical Expert to Medical Expert Communication). Only the authorized medical experts can explore the patients' image because of Kerberos. The proposed work is mainly to restrict the unauthorized access to get the patients'data. So medical image authentication may be achieved without biometric recognition such as finger prints and eye stamps etc. The EPR itself contains the patients' entire history, so after the extraction process Medical expert can able to identify the patient and also the disease information. In future we can embed the EPR inside the medical image after it got encrypted to achieve more security. To increase the authentication, Medical Expert biometric information can be embedded inside the image in the future. Experiments were conducted using more than 500 (512 × 512) image archives in various modalities from the NIH (National Institute of Health) and Aycan sample digital images downloaded from the internet and tests are conducted. Almost in all images with greater than 15000 bits embedding size and got PSNR of 60.4 dB to 78.9 dB with low distortion in received image because of compression, not because of watermarking and average NPCR (Number of Pixels Change Rate) is 98.9 %.

With all the exciting benefits of IoT in healthcare - from mobile applications to wearable and implantable health gadgets-it becomes prominent to ensure that patients, their medical data and the interactions to and from their medical devices are safe and secure. The security and privacy is being breached when the mobile applications are mishandled or tampered by the hackers by performing reverse engineering on the application leading to catastrophic consequences. To combat against these vulnerabilities, there is need to create an awareness of the potential risks of these devices and effective strategies are needed to be implemented to achieve a level of security defense. In this paper, the benefits of healthcare IoT system and the possible vulnerabilities that may result are presented. Also, we propose to develop solutions against these vulnerabilities by protecting mobile applications using obfuscation and return oriented programming techniques. These techniques convert an application into a form which makes difficult for an adversary to interpret or alter the code for illegitimate purpose. The mobile applications use keys to control communication with the implantable medical devices, which need to be protected as they are the critical component for securing communications. Therefore, we also propose access control schemes using white box encryption to make the keys undiscoverable to hackers.

Recent breakthrough in neural network has led to the birth of Convolutional neural network (CNN) which has been found to be very efficient especially in the areas of image recognition and classification. This success is traceable to the availability of large datasets and its capability to learn salient and complex data features which subsequently produce a reusable output model (Fθ). The Fθ are often made available (e.g. on cloud as-a-service) for others (client) to train their data or do transfer learning, however, an adversary can perpetrate a model inversion attack on the model Fθ to recover training data, hence compromising the sensitivity of the model buildup data. This is possible because CNN as a variant of deep neural network does memorize most of its training data during learning. Consequently, this has pose a privacy concern especially when a medical or financial data are used as model buildup data. Existing researches that proffers privacy preserving approach however suffer from significant accuracy degradation and this has left privacy preserving model on a theoretical desk. In this paper, we proposed an ϵ-tuple differential privacy approach that is based on neuron impact factor estimation to preserve privacy of CNN model without significant accuracy degradation. We experiment our approach on two large datasets and the result shows no significant accuracy degradation.

Internet of Things (IoT) is spreading increasingly in different areas of application. Accordingly, IoT also gets deployed in health care including ambient assisted living, telemedicine or medical smart homes. However, IoT also involves risks. Next to increased security issues also safety concerns are occurring. Deploying health care sensors and utilizing medical data causes a high need for IoT architectures free of vulnerabilities in order to identify weak points as early as possible. To address this, we are developing a safety and security analysis approach including a standardized meta model and an IoT safety and security framework comprising a customizable analysis language.

Recently, several health care wearable devices which can intervene in health and collect personal health data have emerged in the medical market. Although health care wearable devices promote the integration of multi-layer medical resources and bring new ways of health applications for users, it is inevitable that some problems will be brought. This is mainly manifested in the safety protection of medical and health data and the protection of user's privacy. From the users' point of view, the irrational use of medical and health data may bring psychological and physical negative effects to users. From the government's perspective, it may be sold by private businesses in the international arena and threaten national security. The most direct precaution against the problem is users' initiative. For better understanding, a research model is designed by the following five aspects: Security knowledge (SK), Security attitude (SAT), Security practice (SP), Security awareness (SAW) and Security conduct (SC). To verify the model, structural equation analysis which is an empirical approach was applied to examine the validity and all the results showed that SK, SAT, SP, SAW and SC are important factors affecting users' data security and privacy protection awareness.

Medical Internet of Things (MIoT) offers innovative solutions to a healthier life, making radical changes in people's lives. Healthcare providers are enabled to continuously and remotely monitor their patients for many medial issues outside hospitals and healthcare providers' offices. MIoT systems and applications lead to increase availability, accessibility, quality and cost-effectiveness of healthcare services. On the other hand, MIoT devices generate a large amount of diverse real-time data, which is highly sensitive. Thus, securing medical data is an essential requirement when developing MIoT architectures. However, the MIoT architectures being developed in the literature have many security issues. To address the challenge of data security in MIoT, the integration of fog computing and MIoT is studied as an emerging and appropriate solution. By data security, it means that medial data is stored in fog nodes and transferred to the cloud in a secure manner to prevent any unauthorized access. In this paper, we propose a design for a secure fog-cloud based architecture for MIoT.

Recently, the home healthcare system has emerged as one of the most useful technology for e-healthcare. Contrary to classical recording methods of patient's medical data, which are, based on paper documents, nowadays all this sensitive data can be managed and forwarded through digital systems. These make possible for both patients and healthcare workers to access medical data or receive remote medical treatment using wireless interfaces whenever and wherever. However, simplifying access to these sensitive and private data can directly put patient's health and life in danger. In this paper, we propose a secure and lightweight biometric-based remote patient authentication scheme using elliptic curve encryption through which two mobile healthcare system communication parties could authenticate each other in public mobile healthcare environments. The security and performance analysis demonstrate that our proposal achieves better security than other concurrent schemes, with lower storage, communication and computation costs.

The panic among medical control, information, and device administrators is due to surmounting number of high-profile attacks on healthcare facilities. This hostile situation is going to lead the health informatics industry to cloud-hoarding of medical data, control flows, and site governance. While different healthcare enterprises opt for cloud-based solutions, it is a matter of time when fog computing environment are formed. Because of major gaps in reported techniques for fog security administration for health data i.e. absence of an overarching certification authority (CA), the security provisioning is one of the the issue that we address in this paper. We propose a security provisioning model (AZSPM) for medical devices in fog environments. We propose that the AZSPM can be build by using atomic security components that are dynamically composed. The verification of authenticity of the atomic components, for trust sake, is performed by calculating the processor clock cycles from service execution at the resident hardware platform. This verification is performed in the fully sand boxed environment. The results of the execution cycles are matched with the service specifications from the manufacturer before forwarding the mobile services to the healthcare cloud-lets. The proposed model is completely novel in the fog computing environments. We aim at building the prototype based on this model in a healthcare information system environment.

Differential privacy is an approach that preserves patient privacy while permitting researchers access to medical data. This paper presents mechanisms proposed to satisfy differential privacy while answering a given workload of range queries. Representing input data as a vector of counts, these methods partition the vector according to relationships between the data and the ranges of the given queries. After partitioning the vector into buckets, the counts of each bucket are estimated privately and split among the bucket's positions to answer the given query set. The performance of the proposed method was evaluated using different workloads over several attributes. The results show that partitioning the vector based on the data can produce more accurate answers, while partitioning the vector based on the given workload improves privacy. This paper's two main contributions are: (1) improving earlier work on partitioning mechanisms by building a greedy algorithm to partition the counts' vector efficiently, and (2) its adaptive algorithm considers the sensitivity of the given queries before providing results.