Press Release on Malware Research: February -2019

DroidCat: Effective mechanical man Malware Detection and Categorization via App-Level identification

Most existing humanoid malware detection and categorization techniques are static approaches, that suffer from evasion attacks, like obfuscation. By analyzing program behaviors, dynamic approaches are doubtless additional resilient against these attacks. however existing dynamic approaches principally depend on characterizing system calls that are subject to system-call obfuscation. This paper presents DroidCat, a unique dynamic app classification technique, to enrich existing approaches. By employing a various set of dynamic options supported methodology calls and inter-component communication (ICC) Intents while not involving permission, app resources, or system calls whereas totally handling reflection, DroidCat achieves superior strength than static approaches in addition as dynamic approaches hoping on system calls. The options were distilled from a activity characterization study of benign versus malicious apps. Through 3 complementary analysis studies with thirty four 343 apps from numerous sources and spanning the past 9 years, we tend to incontestible the soundness of DroidCat in achieving high classification performance and superior accuracy compared with the 2 progressive peer techniques that represent each static and dynamic approaches. Overall, DroidCat achieved ninety seven F1-measure accuracy systematically for classifying apps evolving over the 9 years, police investigation or categorizing malware, 16%–27% over any of the 2 baselines compared. moreover, our experiments with obfuscated benchmarks confirmed higher strength of DroidCat over these baseline techniques. we tend to conjointly investigated the results of varied style choices on DroidCat’s effectiveness and therefore the most significant options for our dynamic classification. we tend to found that features capturing app execution structure like the distribution of methodology calls over user code and libraries are rather more necessary than typical safety features such as sensitive flows. [1]

A Multimodal Deep Learning technique for automaton Malware Detection victimisation numerous options

With the widespread use of smartphones, the quantity of malware has been increasing exponentially. Among good devices, humanoid devices are the foremost targeted devices by malware thanks to their high quality. This paper proposes a completely unique framework for humanoid malware detection. Our framework uses varied sorts of options to replicate the properties of humanoid applications from various aspects, and also the options are refined victimisation our existence-based or similarity-based feature extraction methodology for effective feature illustration on malware detection. Besides, a multimodal deep learning methodology is projected to be used as a malware detection model. This paper is that the initial study of the multimodal deep learning to be employed in the humanoid malware detection. With our detection model, it absolutely was potential to maximise the advantages of encompassing multiple feature sorts. to judge the performance, we have a tendency to dole out varied experiments with a complete of forty one 260 samples. we have a tendency to compared the accuracy of our model therewith of alternative deep neural network models. moreover, we have a tendency to evaluated our framework in varied aspects as well as the potency in model updates, the utility of various options, and our feature illustration methodology. additionally, we have a tendency to compared the performance of our framework with those of alternative existing strategies as well as deep learning-based methods. [2]

Malware Detection Techniques in Security Gateway

Malware threats still become additional refined, and also the amount and motivation of attackers still multiply. the quantity and quality of attack kits have raised, facultative less virtuoso attackers to lease attack code.Security gateways include some half, like firewall, IDS/IPS, mail security, content filtering, traffic shaping and malware detection system.There are several ways for malware detection in security entrance systems primarily based  on some common methods of identification and detection of malicious codes, however choice of the simplest product betting on company security arrange, defence exhaustive arrange, space of business, scale of business and shoppers, experience of  security officers so on, however we would like to own an ideal malware analysis to try and do the simplest protection against malwares and threats we want to style and implement a helpful and economical malware analysis laboratory [3]

Scientists Stop And Search Malware Hidden In Shortened URLs On Twitter

Cyber-criminals are taking advantage of real-world events with high volumes of traffic on Twitter so as to post links to websites that contain malware.

To combat the threat, laptop scientists have created AN intelligent system to spot malicious links disguised in shortened urls on Twitter. they’re going to check the system within the European soccer Championships next summer. The analysis is co-funded by the Engineering and Physical Sciences analysis Council (EPSRC) and also the Economic and Social analysis Council (ESRC). [4]

Enhancing Malware Detection Accuracy through Graph Based Model

Malicious malware may be a serious threat to end-user within the web. Run-time analysis of a program execution behavior is wide accustomed classify malware’s activities particularly once its signature isn’t available. Towards this finish, most of the prevailing run-time malware detection techniques create use of the knowledge out there within the Application Programming Interface decision sequence in Windows platform. This paper suggests a unique malware revealing model supported graph model by capturing system calls throughout the execution of a suspected practicable. The implementation results make sure that the projected decision graph model has higher detection accuracy rate and conjointly solves the measurability drawback once it’s compared to existing strategies. [5]


[1] Cai H, Meng N, Ryder B, Yao D. Droidcat: Effective android malware detection and categorization via app-level profiling. IEEE Transactions on Information Forensics and Security. 2019 Jun;14(6):1455-70. (web link)

[2] Kim T, Kang B, Rho M, Sezer S, Im EG. A Multimodal Deep Learning Method for Android Malware Detection Using Various Features. IEEE Transactions on Information Forensics and Security. 2019 Mar;14(3):773-88. (web link)

[3] Jalinous A. Malware Detection Techniques in Security Gateway. Data Science Letters. 2019 Jan 29;2(1):32-7. (web link)

[4] Scientists Stop And Search Malware Hidden In Shortened URLs On Twitter

September 25, 2015 (web link)

[5] Enhancing Malware Detection Accuracy through Graph Based Model


Research Scholar, Department of Computer Science and Engineering, Pondicherry Engineering College, Puducherry – 605 014, India.


Department of Computer Science and Engineering, Pondicherry Engineering College, Puducherry – 605 014, India. (web link)

Leave a Reply

Your email address will not be published. Required fields are marked *