SETS-NIRAM Neural Repository of Images for Ransomware Analysis and Modeling

About Dataset

This dataset is collected by converting ransomware and benign PE files into images(grayscale and RGB).It contains 500 images of ransomware and 500 images of benign software.The RGB images are obtained by calculating entropy-bigram-markov transition of the bytes present in PE files.The grayscale images are obtained by calculating entropy of PE files and saving it as an image and applying gabor filters on the processed image.The above process was done in an isolated and controlled environment to avoid ransomware propagation and accidental infection.Reverse engineering samples from images is impossible as images are obfuscated by adding noise and helps in defending against adversarial attacks and can be used for research purposes.