Stacked-autoencoder-based model for COVID-19 diagnosis on CT images

With the outbreak of COVID-19, medical imaging such as computed tomography (CT) based diagnosis is proved to be an effective way to fight against the rapid spread of the virus. Therefore, it is important to study computerized models for infectious detection based on CT imaging. New deep learning-based approaches are developed for CT assisted diagnosis of COVID-19. However, most of the current studies are based on a small size dataset of COVID-19 CT images as there are less publicly available datasets for patient privacy reasons. As a result, the performance of deep learning-based detection models needs to be improved based on a small size dataset. In this paper, a stacked autoencoder detector model is proposed to greatly improve the performance of the detection models such as precision rate and recall rate. Firstly, four autoencoders are constructed as the first four layers of the whole stacked autoencoder detector model being developed to extract better features of CT images. Secondly, the four autoencoders are cascaded together and connected to the dense layer and the softmax classifier to constitute the model. Finally, a new classification loss function is constructed by superimposing reconstruction loss to enhance the detection accuracy of the model. The experiment results show that our model is performed well on a small size COVID-2019 CT image dataset. Our model achieves the average accuracy, precision, recall, and F1-score rate of 94.7%, 96.54%, 94.1%, and 94.8%, respectively. The results reflect the ability of our model in discriminating COVID-19 images which might help radiologists in the diagnosis of suspected COVID-19 patients.

     

Prediction of Alzheimer’s Using Random Forest with Radiomic Features

Alzheimer’s disease is a non-reversible, non-curable, and progressive neurological disorder that induces the shrinkage and death of a specific neuronal population associated with memory formation and retention. It is a frequently occurring mental illness that occurs in about 60%–80% of cases of dementia. It is usually observed between people in the age group of 60 years and above. Depending upon the severity of symptoms the patients can be categorized in Cognitive Normal (CN), Mild Cognitive Impairment (MCI) and Alzheimer’s Disease (AD). Alzheimer’s disease is the last phase of the disease where the brain is severely damaged, and the patients are not able to live on their own. Radiomics is an approach to extracting a huge number of features from medical images with the help of data characterization algorithms. Here, 105 number of radiomic features are extracted and used to predict the alzhimer’s. This paper uses Support Vector Machine, K-Nearest Neighbour, Gaussian Naïve Bayes, eXtreme Gradient Boosting (XGBoost) and Random Forest to predict Alzheimer’s disease. The proposed random forest-based approach with the Radiomic features achieved an accuracy of 85%. This proposed approach also achieved 88% accuracy, 88% recall, 88% precision and 87% F1-score for AD vs. CN, it achieved 72% accuracy, 73% recall, 72% precisionand 71% F1-score for AD vs. MCI and it achieved 69% accuracy, 69% recall, 68% precision and 69% F1-score for MCI vs. CN. The comparative analysis shows that the proposed approach performs better than others approaches.     

Privacy preserving security using biometrics in cloud computing

Cloud computing and the efficient storage provide new paradigms and approaches designed at efficiently utilization of resources through computation and many alternatives to guarantee the privacy preservation of individual user. It also ensures the integrity of stored cloud data, and processing of stored data in the various data centers. However, to provide better protection and management of sensitive information (data) are big challenge to maintain the confidentiality and integrity of data in the cloud computation. Thus, there is an urgent need for storing and processing the data in the cloud environment without any information leakage. The sensitive data require the storing and processing mechanism and techniques to assurance the privacy preservation of individual user, to maintain the data integrity, and preserve confidentiality. Face recognition has recently achieved advancements in the unobtrusive recognition of individuals to maintain the privacy-preservation in the cloud computing. This paper emphasizes on cloud security and privacy issues and provides the solution using biometric face recognition. We propose a biometrics face recognition approach for security and privacy preservation of cloud users during their access to cloud resources. The proposed approach has three steps: (1) acquisition of face images (2) preprocessing and extraction of facial feature (3) recognition of individual using encrypted biometric feature. The experimental results establish that our proposed recognition approach can ensure the privacy and security of biometrics data.

     

XAI-driven model for crop recommender system for use in precision agriculture

Agriculture serves as the predominant driver of a country's economy, constituting the largest share of the nation's manpower. Most farmers are facing a problem in choosing the most appropriate crop that can yield better based on the environmental conditions and make profits for them. As a consequence of this, there will be a notable decline in their overall productivity. Precision agriculture has effectively resolved the issues encountered by farmers. Today's farmers may benefit from what's known as precision agriculture. This method takes into account local climate, soil type, and past crop yields to determine which varieties will provide the best results. The explainable artificial intelligence (XAI) technique is used with radial basis functions neural network and spider monkey optimization to classify suitable crops based on the underlying soil and environmental conditions. The XAI technology would provide assets in better transparency of the prediction model on deciding the most suitable crops for their farms, taking into account a variety of geographical and operational criteria. The proposed model is assessed using standard metrics like precision, recall, accuracy, and F1-score. In contrast to other cutting-edge approaches discussed in this study, the model has shown fair performance with approximately 12% better accuracy than the other models considered in the current study. Similarly, precision has improvised by 10%, recall by 11%, and F1-score by 10%.     

Trusted data sharing with flexible access control based on blockchain

Along with the progress of cloud service, a growing quantity of data owners store their data on cloud databases, which can not only reduce data owners’ storage cost but also provide a quick search function. However, while cloud storage brings some conveniences to users, new privacy problems may emerge, such as the leakage of data privacy and user’s query privacy. The best way of protecting data privacy is to encrypt the data. So how to efficiently retrieve the ciphertext to make it available becomes a hot issue in recent years. In this paper, new searchable encryption with multiple keywords is described, it can improve the accuracy of retrieval results, and we present a secure and trusted data sharing framework based on attribute-based encryption (ABE), searchable encryption, and blockchain. Unlike the previous studies, we realize flexible data sharing by using ABE. Furthermore, we transfer the related calculation of ciphertext retrieval to blockchain for credible execution without relying on any trusted third party. The security analysis proves that our method meets the proposed security requirements of data, keyword index, trapdoor, and query. Finally, the experimental results indicate that our scheme suggested has certain practicability and efficiency.     

基于联邦学习的跨源数据错误检测方法

随着海量数据的涌现和不断积累,数据治理成为提高数据质量、最大化数据价值的重要手段.其中,数据错误检测是提高数据质量的关键步骤,近年来引起了学术界及工业界的广泛关注.目前,绝大多数错误检测方法只适用于单数据源场景.然而在现实场景中,数据往往不集中存储与管理.不同来源且高度相关的数据能够提升错误检测的精度.但由于数据隐私安全问题,跨源数据往往不允许集中共享.鉴于此,提出了一种基于联邦学习的跨源数据错误检测方法 FeLeDetect,以在数据隐私保证的前提下,利用跨源数据信息提高错误检测精度.为了充分捕获每一个数据源的数据特征,首先提出一种基于图的错误检测模型GEDM,并在此基础上设计了一种联邦协同训练算法FCTA,以支持在各方数据不出本地的前提下,利用跨源数据协同训练GEDM.此外,为了降低联邦训练的通信开销和人工标注成本,还提出了一系列优化方法.最后,在3个真实数据集上进行了大量的实验.实验结果表明:(1)相较于5种现有最先进的错误检测方法,GEDM在本地场景和集中场景下,错误检测结果的F1分数平均提高了10.3%和25.2%;(2) FeLeDetect错误检测结果的F1分数较本地场景...     

Feature matching and instance reweighting with transfer learning for human activity recognition using smartphone

Human activity recognition using smartphone has been attracting great interest. Since collecting large amount of labeled data is expensive and time-consuming for conventional machine learning techniques, transfer learning techniques have been proposed for activity recognition. However, existing transfer learning techniques typically rely on feature matching based on global domain shift and lack considering the intra-class knowledge transfer. In this paper, a novel transfer learning technique is proposed for cross-domain activity recognition, which can properly integrate feature matching and instance reweighting across the source and target domain in principled dimensionality reduction. The experiments using three real datasets demonstrate that the proposed scheme can achieve much higher precision (92%), recall (91%), and F1-score (92%), in comparison with the existing schemes.

     

大规模差异化点云数据下的联邦语义分割算法

海量点云数据的存储对自动驾驶实时3D协同感知具有重要意义,然而出于数据安全保密性的要求,部分数据拥有者不愿共享其私人的点云数据,限制了模型训练准确性的提升。联邦学习是一种注重数据隐私安全的计算范式,提出了一种基于联邦学习的方法来解决车辆协同感知场景下的大规模点云语义分割问题。融合具有点间角度信息的位置编码方式并对邻近点进行几何衍射处理以增强模型的特征提取能力,最后根据本地模型的生成质量动态调整全局模型的聚合权重,提高数据局部几何结构的保持能力。在SemanticKITTI,SemanticPOSS和Toronto3D三个数据集上进行了实验,结果表明该算法显著优于单一训练数据和基于FedAvg的方法,在充分挖掘点云数据价值的同时兼顾各方数据的隐私敏感性。     

一种基于BloomFilter的改进型加密文本模糊搜索机制研究

随着云计算的日益普及,为实现共享计算资源、节约经济成本等目的,越来越多的重要数据被从本地外包迁移至云端.出于对保护云端数据安全和用户隐私等方面的考虑,数据所用者一般倾向对敏感数据进行加密处理,在此基础上,如何能够对数据开展有效检索处理成为关注的重点.为此,提出一种改进的密文数据多关键字检索机制,一方面,基于BloomFilter数据结构设计一种新的关键字转换方法,能够在保持模糊搜索功能及识别率的同时,有效降低数据索引规模;另一方面,基于动态混淆参数调节的思路改进相似度评估算法,以提高数据的加密强度,并且能很好地反映用户的检索偏好.实验结果验证了所提机制是可行和高效的.     

Efficient Deep Learning Framework for Fire Detection in Complex Surveillance Environment

To prevent economic, social, and ecological damage, fire detection and management at an early stage are significant yet challenging. Although computationally complex networks have been developed, attention has been largely focused on improving accuracy, rather than focusing on real-time fire detection. Hence, in this study, the authors present an efficient fire detection framework termed E-FireNet for real-time detection in a complex surveillance environment. The proposed model architecture is inspired by the VGG16 network, with significant modifications including the entire removal of Block-5 and tweaking of the convolutional layers of Block-4. This results in higher performance with a reduced number of parameters and inference time. Moreover, smaller convolutional kernels are utilized, which are particularly designed to obtain the optimal details from input images, with numerous channels to assist in feature discrimination. In E-FireNet, three steps are involved: preprocessing of collected data, detection of fires using the proposed technique, and, if there is a fire, alarms are generated and transmitted to law enforcement, healthcare, and management departments. Moreover, E-FireNet achieves 0.98 accuracy, 1 precision, 0.99 recall, and 0.99 F1-score. A comprehensive investigation of various Convolutional Neural Network (CNN) models is conducted using the newly created Fire Surveillance SV-Fire dataset. The empirical results and comparison of numerous parameters establish that the proposed model shows convincing performance in terms of accuracy, model size, and execution time.     

Improved Cloud Storage Encryption Using Block Cipher-Based DNA Anti-Codify Model

When it comes to data storage, cloud computing and cloud storage providers play a critical role. The cloud data can be accessed from any location with an internet connection. Additionally, the risk of losing privacy when data is stored in a cloud environment is also increased. A variety of security techniques are employed in the cloud to enhance security. In this paper, we aim at maintaining the privacy of stored data in cloud environment by implementing block-based modelling to boost the privacy level with Anti-Codify Technique (ACoT) and block cipher-based algorithms. Initially, the cipher text is generated using Deoxyribo Nucleic Acid (DNA) model. Block-cipher-based encryption is used by ACoT, but the original encrypted file and its extension are broken up into separate blocks. When the original file is broken up into two separate blocks, it raises the security level and makes it more difficult for outsiders to cloud data access. ACoT improves the security and privacy of cloud storage data. Finally, the fuzzy-based classification is used that stores various access types in servers. The simulation results shows that the ACoT-DNA method achieves higher entropy against various block size with reduced computational cost than existing methods.     

多云存储中保护隐私的数据完整性验证方案

随着云存储的应用,越来越多的用户选择将数据分散地存储在多个云服务器上,但是这种远程存储方式给用户数据的完整性带来了挑战。同时,代替用户校验数据完整性的第三方审计（TPA）近来也被指出存在泄露用户数据隐私的风险。针对现有的远程数据安全性、隐私性及高效验证的问题,提出一种多用户多服务器环境下支持隐私保护的批处理数据完整性验证方案。方案在一般群模型和随机谕言机模型下是可证明安全的。性能分析和实验表明,与其他在多用户多服务器环境下拓展并保护隐私的方案相比,该方案具有较低的通信复杂度和计算复杂度。     

Cooperative private searching in clouds

With the increasing popularity of cloud computing, there is increased motivation to outsource data services to the cloud to save money. An important problem in such an environment is to protect user privacy while querying data from the cloud. To address this problem, researchers have proposed several techniques. However, existing techniques incur heavy computational and bandwidth related costs, which will be unacceptable to users. In this paper, we propose a cooperative private searching (COPS) protocol that provides the same privacy protections as prior protocols, but with much lower overhead. Our protocol allows multiple users to combine their queries to reduce the querying cost while protecting their privacy. Extensive evaluations have been conducted on both analytical models and on a real cloud environment to examine the effectiveness of our protocol. Our simulation results show that the proposed protocol reduces computational costs by 80% and bandwidth cost by 37%, even when only five users query data.     

面向企业数据孤岛的联邦排序学习

排序学习（Learning-to-Rank,LTR）模型在信息检索领域取得了显著成果.而该模型的传统训练方法需要收集大规模文本数据.然而,随着数据隐私保护日渐受到人们重视,从多个数据拥有者（如企业）手中收集数据训练排序学习模型的方式变得不可行.各企业之间数据被迫独立存储,形成了数据孤岛.由于排序模型训练需要使用查询记录、文档等诸多隐私信息,数据孤岛难以融合打通,这制约了排序学习模型的训练.联邦学习能够让多数据拥有方在隐私保护的前提下联合训练模型,是一种打通数据孤岛的新方法.本文在其启发下提出了一种新的框架,即面向企业数据孤岛的联邦排序学习,它同时解决了联邦学习场景下排序学习所面临的两大挑战,即交叉特征生成与缺失标签处理.为了应对多方交叉特征的生成问题,本文使用了一种基于略图（Sketch）数据结构与差分隐私的方法,其相比于传统加密方法具有更高的效率,同时还具有隐私性与结果精度的理论保证.为了应对缺失标签问题,本文提出了一种新的联邦半监督学习方法.最终,本文通过在公开数据集上的大量实验验证了所提方法的有效性.     

An efficient PHR service system supporting fuzzy keyword search and fine-grained access control

Outsourcing of personal health record (PHR) has attracted considerable interest recently. It can not only bring much convenience to patients, it also allows efficient sharing of medical information among researchers. As the medical data in PHR is sensitive, it has to be encrypted before outsourcing. To achieve fine-grained access control over the encrypted PHR data becomes a challenging problem. In this paper, we provide an affirmative solution to this problem. We propose a novel PHR service system which supports efficient searching and fine-grained access control for PHR data in a hybrid cloud environment, where a private cloud is used to assist the user to interact with the public cloud for processing PHR data. In our proposed solution, we make use of attribute-based encryption (ABE) technique to obtain fine-grained access control for PHR data. In order to protect the privacy of PHR owners, our ABE is anonymous. That is, it can hide the access policy information in ciphertexts. Meanwhile, our solution can also allow efficient fuzzy search over PHR data, which can greatly improve the system usability. We also provide security analysis to show that the proposed solution is secure and privacy-preserving. The experimental results demonstrate the efficiency of the proposed scheme.     

ABKS-SKGA: Attribute-based keyword search secure against keyword guessing attack

In order to guarantee security and privacy of sensitive data, attribute-based keyword search (ABKS) enables data owners to upload their encrypted data to cloud servers, and authorizes intended data users to retrieve it. Meanwhile, ABKS outsources heavy search work to cloud servers, which makes ABKS adaptive to mobile computing environment. However, as cloud servers can both generate keyword ciphertexts and run search algorithm, the existing most ABKS schemes are vulnerable to keyword guessing attack. In this paper, we show the fundamental cause that the existing ABKS schemes do not resist keyword guessing attack is any entity can generate keyword ciphertext. To solve the above problem, in the phase of keyword ciphertext generation, we use private key of data owner to sign keyword prior to generating keyword ciphertext. Therefore, any other entity does not forge keyword ciphertext, which can resist keyword guessing attack. We give the formal definition and security model of attributed-based keyword search secure against keyword guessing attack (ABKS-SKGA). Furthermore, we provide an ABKS-SKGA scheme. The ABKS-SKGA scheme is proved secure against chosen-plaintext attack (CPA). Performance analysis shows that the proposed scheme is practical.     

Achieving fully privacy-preserving private range queries over outsourced cloud data

With the prevalence of cloud computing, data owners are motivated to outsource their databases to the cloud server. However, to preserve data privacy, sensitive private data have to be encrypted before outsourcing, which makes data utilization a very challenging task. Existing work either focus on keyword searches and single-dimensional range query, or suffer from inadequate security guarantees and inefficiency. In this paper, we consider the problem of multidimensional private range queries over encrypted cloud data. To solve the problem, we systematically establish a set of privacy requirements for multidimensional private range queries, and propose a multidimensional private range query (MPRQ) framework based on private block retrieval (PBR), in which data owners keep the query private from the cloud server. To achieve both efficiency and privacy goals, we present an efficient and fully privacy-preserving private range query (PPRQ) protocol by using batch codes and multiplication avoiding technique. To our best knowledge, PPRQ is the first to protect the query, access pattern and single-dimensional privacy simultaneously while achieving efficient range queries. Moreover, PPRQ is secure in the sense of cryptography against semi-honest adversaries. Experiments on real-world datasets show that the computation and communication overhead of PPRQ is modest.     

Approaches and challenges of privacy preserving search over encrypted data

More and more data owners are encouraged to outsource their data onto cloud servers for reducing infrastructure, maintenance cost and also to get ubiquitous access to their stored data. However, security is one issue that discourages data owners from adopting cloud servers for data storage. Searchable Encryption (SE) is one of the few ways of assuring privacy and confidentiality of such data by storing them in encrypted form at the cloud servers. SE enables the data owners and users to search over encrypted data through trapdoors. Most of the user information requirements are fulfilled either through Boolean or Ranked search approaches. This paper aims at understanding how the confidentiality and privacy of information can be guaranteed while processing single and multi-keyword queries over encrypted data using Boolean and Ranked search approaches. This paper presents all possible leakages that happen in SE and also specifies which privacy preserving approach to be adopted in SE schemes to prevent those leakages to help the practitioners and researchers to design and implement secure searchable encryption systems. It also highlights various application scenarios where SE could be utilized. This paper also explores the research challenges and open problems that need to be focused in future.     

MDEV Model: A Novel Ensemble-Based Transfer Learning Approach for Pneumonia Classification Using CXR Images

Pneumonia is a dangerous respiratory disease due to which breathing becomes incredibly difficult and painful; thus, catching it early is crucial. Medical physicians’ time is limited in outdoor situations due to many patients; therefore, automated systems can be a rescue. The input images from the X-ray equipment are also highly unpredictable due to variances in radiologists’ experience. Therefore, radiologists require an automated system that can swiftly and accurately detect pneumonic lungs from chest x-rays. In medical classifications, deep convolution neural networks are commonly used. This research aims to use deep pre-trained transfer learning models to accurately categorize CXR images into binary classes, i.e., Normal and Pneumonia. The MDEV is a proposed novel ensemble approach that concatenates four heterogeneous transfer learning models: MobileNet, DenseNet-201, EfficientNet-B0, and VGG-16, which have been finetuned and trained on 5,856 CXR images. The evaluation matrices used in this research to contrast different deep transfer learning architectures include precision, accuracy, recall, AUC-roc, and f1-score. The model effectively decreases training loss while increasing accuracy. The findings conclude that the proposed MDEV model outperformed cutting-edge deep transfer learning models and obtains an overall precision of 92.26%, an accuracy of 92.15%, a recall of 90.90%, an auc-roc score of 90.9%, and f-score of 91.49% with minimal data pre-processing, data augmentation, finetuning and hyperparameter adjustment in classifying Normal and Pneumonia chests.     

Using a deep learning network to recognise low back pain in static standing

Abstract

Low back pain (LBP) remains one of the most prevalent musculoskeletal disorders, while algorithms that able to recognise LBP patients from healthy population using balance performance data are rarely seen. In this study, human balance and body sway performance during standing trials were utilised to recognise chronic LBP populations using deep neural networks. To be specific, 44 chronic LBP and healthy individuals performed static standing tasks, while their spine kinematics and centre of pressure were recorded. A deep learning network with long short-term memory units was used for training, prediction and implementation. The performance of the model was evaluated by: (a) overall accuracy, (b) precision, (c) recall, (d) F1 measure, (e) receiver-operating characteristic and (f) area under the curve. Results indicated that deep neural networks could recognise LBP populations with precision up to 97.2% and recall up to 97.2%. Meanwhile, the results showed that the model with the C7 sensor output performed the best.

Practitioner summary: Low back pain (LBP) remains the most common musculoskeletal disorder. In this study, we investigated the feasibility of applying artificial intelligent deep neural network in detecting LBP population from healthy controls with their kinematics data. Results showed a deep learning network can solve the above classification problem with both promising precision and recall performance.