(i) Used the raw malware executable files to extract grayscale representations of malicious characteristics (ii) By extracting visual elements from such photos, malware can be analyzed.
(i) Worked with a malware picture dataset that included 9,342 malware samples from 25 distinct types. (ii) They were the first to examine the use of byte graphs as grayscale pictures for automated malware categorization.
(i) Utilized the approach of [20] to extract GIST characteristics from grayscale pictures and categorize them using the Euclidean distance as a metric. (ii) However, their method has a considerable computational cost.
(i) Built an effective texture-based feature vector from the malware images using the wavelet transform. (ii) Conducted malware classification using a multiclass support vector machine with malware input as images.
(i) Performed a simple CNN detected the variant of codes by turning them into grayscale images. Kalash et al. [25] classified malware by using two datasets, Malimg [15] and Microsoft [26]. (ii) Converted the malware binaries into malware images, their approach achieved high accuracy of 98.52% and 99.97%.
(i) Used two different approaches of feature extractions to classify Windows API Calls database. (ii) Shows that they depend on inverse document frequency vector and categorical vector. (iii) Proposed method score high accuracy above 90.0%.
