Overview

A web-based application built with Flask for detecting and classifying suspicious binary files as malware. The system uses a computer vision-based approach with a deep learning model (ResNet-RS50) trained on the MaleVis dataset. The model used in this system achieves an accuracy of approximately 86.91%, trained on the MaleVis dataset which consists of 9,100 samples from 25 malware families. Since the dataset only includes a limited number of known families, the system may misclassify files belonging to unknown or novel malware types.

Repository

Tech Stacks