Creation of a heterogeneous video dataset for forensic analyses

DegreeBachelor / Master
StatusAvailable
Supervisor(s)Verena Lachner, MSc

Description

Digital forensics deals with the scientific reconstruction of digital traces for use in a court of law. Given that more than 80% of internet traffic is video, digital video forensics is an area of growing research interest. However, the structural dependencies in encoded video streams are not yet fully understood, hindering advances in this field. In addition, the use of different encoder implementations leads to different results, which can affect the effectiveness of forensic methods. These results can also help to identify the specific encoder implementation used to compress a video. There is a need for a systematic analysis of encoding decisions and parameters.

The objective of a Bachelor’s thesis is to create a heterogeneous video dataset, which can serve as a foundation for future analysis tasks. This involves the production (i.e., selection or creation) of videos with diverse processing histories, encompassing factors such as double compression and compression parameters. The student has to make sure that differences in the encoding pipeline also appear in the video stream. For this task we provide a simple analysis library in Julia. In the thesis, the student documents their approach, assesses the quality of the dataset, and offers recommendations for its maintenance. If this topic is taken for a Master’s thesis, we also expect that the student is able to use these insights to identify and measure differences/commonalities between different encoder implementations and propose a classifier to distinguish sources.

References

  • Wiegand, T., Sullivan, G.J., Bjontegaard, G., and Luthra, A. Overview of the H.264/AVC Video Coding Standard. IEEE Transactions on Circuits and Systems for Video Technology, 13, 7 (2003), 560–576.
  • Lachner, V., Schaar, K., and Zimmermann, R. CSM in Motion Vector Steganalysis: The Effect of Coders on Motion Vectors in H.264 Video Encoding. In IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP). Rhodos, Greece, 2023. [Publisher]
  • Böhme, R. and Westfeld, A. Statistical Characterisation of MP3 Encoders for Steganalysis. In ACM Multimedia and Security Workshop (MMSEC). ACM Press, New York, 2004, pp. 25–34. [Publisher]