Paper Title
Detecting Windows Malware Using Image Analyzing for Extracted Behavior Features

Abstract
Malware detection is pivotal for safeguarding digital ecosystems, relying on static and dynamic analyses and recently artificial intelligence has been used as tool in this field. The Artificial Intelligence has been used to analysis the malware itself, to analysis the malware’s behavior, to classify the malware families and also to detect malwares . This research integrates dynamic malware analysis and AI to detect malware. This research focuses on behavioral analysis, extracting ten features from isolated virtual machines, such as process ID, name, user, CPU percentage, network connections, and more. The dataset comprises 50 malware samples and 11 benign programs. Text-based classification, utilizing feedforward neural networks and recurrent neural networks (RNN) achieves accuracy rates of 56% and 68%, respectively. Additionally, the extracted features are transformed into grayscale images for image-based classification, employing a CNN with a resulting accuracy rate of 70.1%. Static and dynamic analyses' strengths and limitations are discussed. Keywords - Malware Analysis, Dynamic Analysis, Image classification, Malware Behavior extraction, Text Classification