Abstract Knee disorders are common among the human population. Knee osteoarthritis (OA) is the most widespread knee joint disorder, which may require surgical treatment. The detection and diagnosis of knee joint disorders from medical images demand enormous human effort and time. The development of a computer‐aided diagnosis (CAD) system can notably minimise the burden of medical experts and remove the intra‐observer and inter‐observer variations. To achieve the goal, the highly challenging research problem of knee image segmentation has been frequently paid attention in past years, which can be efficiently applied in the development of the CAD system. Knee image segmentation is a challenging task owing to the image contrasts, intensity variations, shape irregularities, and the presence of thin cartilage structures. Therefore, this paper presents a literature review of automated segmentation approaches mainly focused on the segmentation of knee cartilage and bone, with respect to the underlying technical aspects, datasets used, and the performance reported. The paper also presents the growth from classical segmentation approaches towards the deep learning approaches in the knee image segmentation. Owing to the varying quality and complexity of different knee image datasets, this paper abstains from doing a rigorous comparative evaluation of image segmentation approaches.

In the present digital world, usage of the internet has increased many folds as users have become dependent on the cloud-based applications. The disclosure of personal information on such platforms becomes a prospective threat for an attack. Researchers have used randomised data distortion technique with addition of random noise to conceal the sensitive data from an unauthorised adversary. This perturbation technique has relevance for the numerical datasets only. In this paper, we propose a hybrid model of two phases encoding with additive random noise value for ensuring non-disclosure of private and sensitive information and maintaining an effective balance between data privacy and data utility. The proposed technique has been tested on different data sizes of the real estate industry in terms of efficiency and effectiveness in preserving privacy and data utility. The proposed algorithm has been evaluated in terms of privacy level and information loss. It has proved effective in comparison with other privacy-preserving techniques such as perturbation and encryption in terms of space complexity and efficiency.

With the use of identity resolution, both information leakage and identity hacking can be reduced to some extent. In this paper, a prototype has been developed to classify Twitter users as suspicious and nonsuspicious on the basis of features which identify user demographics and their tweeting activity using Twitter APIs. A model has been devised based upon user and tweet meta-data which is used to calculate user score and tweet score, and further aggregate the values generated by these scores to label suspicious and nonsuspicious users in the collected dataset of around 21,492 Twitter users. Further, support vector machine classifier has been used to classify the labeled data. Through this paper, our analysis about the role of features and the characteristics of dataset used for the categorization of users in Twitter has been reported. The experimental results illustrate that the proposed system can identify suspicious users with an accuracy of 94.1%.

Social networks once being an innoxious platform for sharing pictures and thoughts among a small online community of friends has now transformed into a powerful tool of information, activism, mobilization, and sometimes abuse. Detecting true identity of social network users is an essential step for building social media an efficient channel of communication. This paper targets the microblogging service, Twitter, as the social network of choice for investigation. It has been observed that dissipation of pornographic content and promotion of followers market are actively operational on Twitter. This clearly indicates loopholes in the Twitter’s spam detection techniques. Through this work, five types of spammers-sole spammers, pornographic users, followers market merchants, fake, and compromised profiles have been identified. For the detection purpose, data of around 1 Lakh Twitter users with their 20 million tweets has been collected. Users have been classified based on trust, user and content based features using machine learning techniques such as Bayes Net, Logistic Regression, J48, Random Forest, and AdaBoostM1. The experimental results show that Random Forest classifier is able to predict spammers with an accuracy of 92.1%. Based on these initial classification results, a novel system for real-time streaming of users for spam detection has been developed. We envision that such a system should provide an indication to Twitter users about the identity of users in real-time.

In the arena of internet of things, everyone has the ability to share every aspect of their lives with other people. Social media is the most popular and effective medium to provide communication. Social media has gripped our lives in a dramatic way. Privacy of users data lying with the service providers needs to be preserved when published for the purpose of research as the release of sensitive personal information of an individual may pose security threats. This has become an important research area nowadays. To some extent, the concepts of anonymization that were earlier used to preserve privacy of relational microdata have been applied to preserve privacy of social networks data. Anonymizing social networks data is challenging as it is a complex structure with users connected to one another graphically and the most important is to preserve the structural properties of the graph depicting the social network relationships while applying such concepts. Recent studies based upon K-anonymity and L-diversity help to preserve privacy of online social networks data and subsequently identify attacks that arise while applying these techniques in different scenarios. K-anonymity equalizes the degree of the nodes to prevent the data from identity disclosure but it cannot preserve sensitive information and also cannot handle attacks arising due to background knowledge and homogeneity. To cope up with the drawbacks of K anonymity, L-diversity was introduced that protects the sensitive labels of the users. In this paper, a novel technique has been proposed which implements the combined features of K-anonymity and L-diversity. Our proposed approach has been validated using the data of real time social network–Twitter (most popular microblogging network). The performance of the proposed technique has been measured by the metrics, such as average path length, average change in sensitive labels, and remaining ratio of top influential users. It thus becomes evident from the results that the values of these parameters attained with the proposed technique for the anonymized graph has minimal variation to that of original structural graph. So, it is possible to retain the utility without compromising privacy while publishing social networks data. Further, the performance of the proposed technique has been discussed by calculating the information loss that addresses the concern of preserving privacy with the least variation of actual content viz info loss.

Academicians and industrialists working on malware use static and dynamic analysis in order to understand their functionality and the menace level posed by them. Industries providing anti-malware solutions calculate malware threat level using the approaches that involve human intervention and demand the skilled analysts along with a large number of resources. With the increasing volume, velocity, and complexity of malware, assigning such a large number of resources is not possible. Thus, there is a need to develop techniques that can automatically compute the threat or damage posed by a piece of malware (to a victim machine) as soon as it appears in the wild. This assessment of damage capability level to a zero-day malware can help in providing early warnings about a specific piece of malware so that immediate attention could be paid to it in terms of allocating resources for performing a closer analysis. This paper presents an automated technique based on fuzzy modeling for computing damage potential of malicious programs, which is calculated on the basis of features obtained after performing automated analysis of malware binaries in the sandboxed environment.

Lung field defines a region‐of‐interest in which specific radiologic signs such as septal lines, pulmonary opacities, cavities, consolidations, and lung nodules are searched by a chest radiographic computer‐aided diagnostic system. Thus, its precise segmentation is extremely important. To precisely segment it, numerous methods have been developed during the last four decades. However, no exclusive survey consolidating the advancements in these methods has been presented till date, thus indicating a void and the need. This study fills the void by presenting a comprehensive survey of these methods with a focus on their underlying principle, the dataset used, reported performance, and relative merits and demerits. It refrains from doing a hard comparative evaluation by bringing all of them on a common platform, since the datasets used in their development and testing are of varied quality, complexity, and are not publicly available. It also provides a glimpse of deep learning, the present state of deep‐learning‐based lung field segmentation methods, expectations from it, and the challenges ahead of it.

Online Social Networks act as a popular forum to promote and manage the reputation of aristocrats. Generally, big organizations, politicians, celebrities, and journalists require a huge number of followers/fans to promote and manage their reputation on Online Social Networks such as Twitter. This demand of more followers has originated Twitter Followers Market that deals with the sale and purchase of fake/compromised Twitter accounts. In this paper, an analysis of merchants of this marketing industry has been conducted using graph- and content-based features. The present study has been conducted by collecting 15,750 tweets related to the sale or purchase of Twitter followers posted by around 995 users. The analysis infers that the merchants of this marketing industry fulfill the characteristics of spammers as stated by the rules and policies of Twitter. Further, machine learning classification approach has been used to categorize these merchants from genuine users based on their behavioral features. To the best of our knowledge, this is the novel study to analyze and then categorize the behavior of such merchants involved in the sale of Twitter accounts as spammers.

Social networks like Facebook, Twitter, Pinterest etc. provide data of its users to the demanding organizations to better comprehend the quality of their potential clients. Publishing confidential data of social network users in its raw form raises several privacy and security concerns. Recently, some anonymization techniques have been developed to address these issues. In this paper, a technique to prevent identity disclosure through structure attacks has been proposed which not only prevents identity disclosure but also preserves utility of data published by online social networks. Algorithms have been developed by using noise nodes/edges with the consideration of introducing minimum change in the original graphical structure of social networks. The authors' work is unique in the sense that previous works are based on edge editing only but their proposed work protects against structure attacks using mutual nodes in the social network and the effectiveness of the proposed technique has been proved using APL (Average Path Length) and information loss as parameters.

Malware is one of the most terrible and major security threats facing the Internet today. It is evolving, becoming more sophisticated and using new ways to target computers and mobile devices. The traditional defences like antivirus softwares typically rely on signature based methods and are unable to detect previously unseen malwares. Machine learning approaches have been adopted to classify malwares based on the features extracted using static or dynamic analysis. Both type of malware analysis have their pros and cons. In this paper, we propose a classification framework which uses integration of both static and dynamic features for distinguishing malwares from clean files. A real world corpus of recent malwares is used to validate the proposed approach. The experimental results, based on a dataset of 998 malwares and 428 cleanware files provide an accuracy of 99.58% indicating that the hybrid approach enhances the accuracy rate of malware detection and classification over the results obtained when these features are considered separately.

Malware is one of the most terrible and major security threats facing the Internet today. Anti-malware vendors are challenged to identify, classify and counter new malwares due to the obfuscation techniques being used by malware authors. In this paper, we present a simple, fast and scalable method of differentiating malwares from cleanwares on the basis of features extracted from Windows PE files. The features used in this work are Suspicious Section Count and Function Call Frequency. After automatically extracting features of executables, we use machine learning algorithms available in WEKA library to classify them into malwares and cleanwares. Our experimental results provide an accuracy of over 98% for a data set of 3,087 executable files including 2,460 malwares and 627 cleanwares. Based on the results obtained, we conclude that the Function Call Frequency feature derived from the static analysis method plays a significant role in distinguishing malware files from benign ones.

With the increase in online publishing of social network data the requirement to protect confidential information related to users has become the main concern of publishers. To cater to this need many anonymization techniques like K-anonymity, L-diversity and T-closeness has been proposed by various researchers for micro-data as well as social network data. In this paper we aim to protect sensitive information of users of Twitter-second most popular social networking site. For the purpose of carrying out anonymization, a crawler has been developed to collect data of around 10K users from publicly available information. Data of around 30 users have been used to carry out the experimental work using ARX tool. All three anonymization methods: K-anonymity, L-diversity and T-closeness have been used. Performance of technique is evaluated using information gain as a metric.

The web has become a vital global platform that binds together almost all daily activities like communication, sharing, and collaboration. Impersonators, phishers, scammers and spammers crop up all the time in Online Social Networks (OSNs), and are even harder to identify. People in the public eyes like politicians, celebrities, sports persons, media persons and other public figures with huge followings are particularly vulnerable to this type of attacks. The main objective in this work is to identify those forged users who harm genuine ones, jeopardize the identity and hence the security and privacy of users. In this paper a framework for the detection of malicious users, non-malicious users and celebrities has been developed by using an attribute set for user classification based on user characteristics. For the purpose of detecting malicious users, non-malicious users and celebrities, a crawler has been developed for Twitter and data of around 22K users have been collected from publicly available information. Data of around 7,500 users have been used for training and testing purpose in Weka for classification of users. 5 classifiers have been used and compared on the basis of performance metrics like precision, recall, F-measure and accuracy. RandomForest outperforms all the classifiers with 99.8% accuracy.