Texture Based Malware Pattern Identification and Classification

Abstract

Malware texture pattern plays an essential role in defense against malicious instructions which were analyzed by malware analyst. It is identified as a security threat. Classifying malware samples based on static analysis which is a challenging task. This paper introduces an approach to classify malware variants as a gray scale image based on texture features such as different patterns of malware samples. Malicious samples are classified through the machine learning techniques. The proposed method experimented on malware dataset which is consisting of large number of malware samples. The similarities are calculated by texture analysis methods with Euclidian distance for various variants of malware families. The available samples are named by the Antivirus companies which can analyze through supervised learning techniques. The experimental results show that the effective identification of malware texture pattern through the image processing which gives better accuracy results compared to existing work.