What is it about?
This paper introduces the first physical adversarial attack targeting DL-based compression in the context of live video streaming. The proposed attack, namely CamHack, places a small-sized physical patch in the visual scene covered by the live camera to manipulate the bitrate of compressed content and disrupt live streaming. Extensive experiments in various streaming scenes and network conditions show that CamHack increases bit consumption by 276.58%, 384.74%, and 942.28% over the clean scene without a patch on three representative victim models.
Featured Image
Photo by Swello on Unsplash
Why is it important?
Modern video streaming relies increasingly on DL-based compression methods, , which offers impressive performance but also introduces new security risks, particularly in live streaming. This work identifies a new attack surface: adversaries can now interfere physically with the input scene to exploit the video compression process itself. The propsed attack will increase the bitrate of compressed output, introduce playback rebuffering and stalls, and eventually compromise the viewing experience in live streaming.
Read the Original
This page is a summary of: A Patch Can Disrupt Live Video Streaming: Physical Adversarial Attacks on Deep Learning Compression, ACM Transactions on Multimedia Computing Communications and Applications, July 2025, ACM (Association for Computing Machinery),
DOI: 10.1145/3750047.
You can read the full text:
Contributors
The following have contributed to this page
