![](bg9.png)
4 CONCLUSION AND FUTURE
WORK
In this paper, we proposed an intrusion detection sys-
tem aimed to automatically discriminate between le-
gitimate and malicious network traces. In detail, we
propose to represent network traces in terms of im-
ages to input several deep-learning models to detect
the application that generated the specific network
trace. We take into account also prediction explain-
ability, by adopting the Grad-CAM, aimed to high-
light with the heatmap the areas of the image symp-
tomatic of a certain prediction. The deployment of
two versions of SSIM metric to measure heatmap sim-
ilarity. These metrics help assess the degree of resem-
blance between heatmaps, offering valuable insights
into the effectiveness of various approaches or tech-
niques. The limitations of the study include the use of
a limited dataset and the need for further evaluation of
the model’s robustness with additional algorithms for
heatmap generation. For this we plan as future work
to consider more algorithms for heatmap generation,
in order to evaluate the model’s robustness. More-
over, different activation maps will be considered and
we will explore the possibility to detect also malware
in the IoT environment with the proposed method.
ACKNOWLEDGEMENTS
This work has been partially supported by EU DUCA,
EU CyberSecPro, and EU E-CORRIDOR projects
and PNRR SERICS SPOKE1 DISE, RdS 2022-2024
cybersecurity.
This work has been carried out within the Ital-
ian National Doctorate on Artificial Intelligence run
by the Sapienza University of Rome in collaboration
with the Institute of Informatics and Telematics (IIT),
the National Research Council of Italy (CNR).
REFERENCES
Casolare, R., De Dominicis, C., Iadarola, G., Martinelli, F.,
Mercaldo, F., and Santone, A. (2021). Dynamic mo-
bile malware detection through system call-based im-
age representation. J. Wirel. Mob. Networks Ubiqui-
tous Comput. Dependable Appl., 12(1):44–63.
Dhote, Y., Agrawal, S., and Deen, A. J. (2015). A survey on
feature selection techniques for internet traffic classi-
fication. In 2015 International Conference on Com-
putational Intelligence and Communication Networks
(CICN), pages 1375–1380. IEEE.
Finsterbusch, M., Richter, C., Rocha, E., Muller, J.-A., and
Hanssgen, K. (2013). A survey of payload-based traf-
fic classification approaches. IEEE Communications
Surveys & Tutorials, 16(2):1135–1156.
Goel, R., Mehmood, I., and Ugail, H. (2021). A study of
deep learning-based face recognition models for sib-
ling identification. Sensors, 21(15):5068.
Iadarola, G., Mercaldo, F., Santone, A., and Martinelli, F.
(2022). Assessing deep learning prediction in image-
based malware detection with activation maps. In
Proceedings of the 27th European Symposium on Re-
search in Computer Security (ESORICS).
LeCun, Y., Jackel, L. D., Bottou, L., Cortes, C., Denker,
J. S., Drucker, H., Guyon, I., Muller, U. A., Sackinger,
E., Simard, P., et al. (1995). Learning algorithms
for classification: A comparison on handwritten digit
recognition. Neural networks: the statistical mechan-
ics perspective, 261(276):2.
Li, Z., Qin, Z., Huang, K., Yang, X., and Ye, S. (2017). In-
trusion detection using convolutional neural networks
for representation learning. In Neural Information
Processing: 24th International Conference, ICONIP
2017, Guangzhou, China, November 14–18, 2017,
Proceedings, Part V, pages 858–866. Springer.
Sara, U., Akter, M., and Uddin, M. S. (2019). Image qual-
ity assessment through fsim, ssim, mse and psnr—a
comparative study. Journal of Computer and Commu-
nications, 7(3):8–18.
Selvaraju, R. R., Das, A., Vedantam, R., Cogswell, M.,
Parikh, D., and Batra, D. (2016). Grad-cam: Why
did you say that? arXiv preprint arXiv:1611.07450.
Sharma, N., Jain, V., and Mishra, A. (2018). An analysis
of convolutional neural networks for image classifica-
tion. Procedia computer science, 132:377–384.
University, C. (2016). The stratosphere ips project
dataset. https://www.stratosphereips.org/category/
dataset.html.
Vinayakumar, R., Soman, K., and Poornachandran, P.
(2017). Applying convolutional neural network for
network intrusion detection. In 2017 International
Conference on Advances in Computing, Communica-
tions and Informatics (ICACCI), pages 1222–1228.
IEEE.
Wang, W., Zhu, M., Zeng, X., Ye, X., and Sheng, Y.
(2017). Malware traffic classification using convo-
lutional neural network for representation learning.
In 2017 International conference on information net-
working (ICOIN), pages 712–717. IEEE.
Wei Wang, D. L. (2016). Ustc-tfc2016. https://github.com/
yungshenglu/USTC-TFC2016.
A Method for Robust and Explainable Image-Based Network Traffic Classification with Deep Learning
393