Malware clustering using image processing hashes

Malware clustering is a big challenge and because of the large amount of samples, it is a problem in Big Data processing. Since the number of Malwares is huge, general clustering algorithms are not applicable to Malware clustering for its time complexity. The time complexity is high since we need to compute similarity between every pair of samples that costs O (n²). In this paper, we apply image processing perceptual hashing methods that do not need any comparison between samples and are really fast and applicable to a huge number of samples. Comparison results of our method to other methods indicate better scalability and good accuracy.