bar371/malware-detection

Malware-Detection Using Transfer Learning of ResNet-50

• Implementation of the article "Malicious Software Classification using Transfer Learning of ResNet-50 Deep Neural Network" under "classifier.ipynb". Achieved 98% Accuracy.

• Malware detection using transfer learning from ResNet-50, with only 2 classes- benign and malicious, i.e- distinction between benign and malicious software.

• Data:

  • MalImg dataset- consists of 9,339 labeled malware binary images in 25 different malware classes.
  • Benign dataset- created by us, consists of 7592 benign binary images.