I am working on malware analysis. I know that publically available data set Malimg. I woulde like to know is there any data set publically available for research purpose.
You can get it from " A Signal Processing Approach To Malware Analysis"
https://www.alexandria.ucsb.edu/lib/ark:/48907/f3f47npp
I am also working on Android Malware detection. I just started any advise for me. Thanks
I used this repository: https://github.com/ytisf/theZoo
for malware analysis
Conference Paper MalViz: An Interactive Visualization Tool for Tracing Malware
The above mentioned link contains only a single sample for each most of the malware family.
Machine learning requires more samples for each malware family
Hi,
I have recently published the "MaleVis" dataset involving 25 malware + legitimate software binary images (9100 training + 4660 testing) . The images we have extracted have either 224x224 or 300x300 resolutions. It has been evaluated with CNNs and DenseNet has 97.48% accuracy as SOTA. The MaleVis dataset is freely available at https://web.cs.hacettepe.edu.tr/~selman/malevis/
Well, the language of the conference is Turkish. So, currently there is no English based one. The abstract is in English. However, you can ask me your questions and I can reply. Note that, MaleVis can be used for academic purposes freely.
You can use the Microsoft Big 2015 dataset which is available on kaggle. The code to transform a byte file to a grayscale image would be straightforward. If I remember, the winning solution has used a transformation to a gray scale image and use just the first pixels of the image as one of their "golden features" to achieve a very high accuracy. I used this dataset in my research. Good luck!
Thank you. If you have the Microsoft Big 2015 dataset in the form of gray scale image, could you please share it.
https://web.cs.hacettepe.edu.tr/~selman/malevis/bin2png.py
I have modified the original Sultanik's code in order to be run under python3. With this code, you can convert binary files to rgb images. As I stated in malevis paper, the RGB based 3 channel images also work well. However, if you need grayscale only images, then you must change the script I have given above.
For a couple of weeks, we are planning to transform Microsoft's data set into both grayscale and rgb images. Stay tuned.
- Badges
- Science topic
- Similar topics
- Computer Science
- Images