| Title | A New Method for Ransomware Detection Based on PE Header Using Convolutional Neural Networks |
| Publication Type | Conference Paper |
| Year of Publication | 2020 |
| Authors | Manavi, Farnoush, Hamzeh, Ali |
| Conference Name | 2020 17th International ISC Conference on Information Security and Cryptology (ISCISC) |
| Date Published | sep |
| Keywords | Byte, composability, convolution neural network, convolutional neural networks, cryptography, Feature, feature extraction, image, Information security, Metrics, Neural networks, Operating systems, pubcrawl, ransomware, ransomware detection, Resiliency |
| Abstract | With the spread of information technology in human life, data protection is a critical task. On the other hand, malicious programs are developed, which can manipulate sensitive and critical data and restrict access to this data. Ransomware is an example of such a malicious program that encrypts data, restricts users' access to the system or their data, and then request a ransom payment. Many types of research have been proposed for ransomware detection. Most of these methods attempt to identify ransomware by relying on program behavior during execution. The main weakness of these methods is that it is not clear how long the program should be monitored to show its real behavior. Therefore, sometimes, these researches cannot early detect ransomware. In this paper, a new method for ransomware detection is proposed that does not require running the program and uses the PE header of the executable files. To extract effective features from the PE header files, an image based on PE header is constructed. Then, according to the advantages of Convolutional Neural Networks in extracting features from images and classifying them, CNN is used. The proposed method achieves 93.33% accuracy. Our results indicate the usefulness and practicality method for ransomware detection. |
| DOI | 10.1109/ISCISC51277.2020.9261903 |
| Citation Key | manavi_new_2020 |
A New Method for Ransomware Detection Based on PE Header Using Convolutional Neural Networks