Toward Accessible and Safe Live Streaming Using Distributed Content Filtering with MoQ

2025-05-13

AI Summary

To address the challenge of balancing low latency and high safety in real-time live video content moderation, this paper proposes an end-to-end dynamic filtering framework tailored for social platforms. The method decentralizes content filtering to clients for the first time, leveraging an extended Media over QUIC (MoQ) protocol to enable GOP-level streaming policy enforcement, collaborative client-side analysis, and lightweight photosensitive seizure detection. The system selectively removes only non-compliant segments while immediately resuming playback—ensuring safety for photosensitive users with an end-to-end latency of only 200–500 ms (i.e., one GOP duration). Experimental results demonstrate substantial improvements over conventional centralized moderation: the framework achieves breakthroughs in accessibility, real-time performance, and safety, establishing a new trade-off frontier between latency and security in live streaming moderation.

Technology Category

• Computer Vision: Multi-modal Vision
• Natural Language Processing: Safety and Robustness
• Machine Learning: Multimodal Learning

Application Category

• Security and Privacy: Large-scale security measurements
• Responsible Web: Consent frameworks and practices on the web
• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG

Abstract

Live video streaming is increasingly popular on social media platforms. With the growth of live streaming comes an increased need for robust content moderation to remove dangerous, illegal, or otherwise objectionable content. Whereas video on demand distribution enables offline content analysis, live streaming imposes restrictions on latency for both analysis and distribution. In this paper, we present extensions to the in-progress Media Over QUIC Transport protocol that enable real-time content moderation in one-to-many video live streams. Importantly, our solution removes only the video segments that contain objectionable content, allowing playback resumption as soon as the stream conforms to content policies again. Content analysis tasks may be transparently distributed to arbitrary client devices. We implement and evaluate our system in the context of light strobe removal for photosensitive viewers, finding that streaming clients experience an increased latency of only one group-of-pictures duration.

Problem

• Real-time content moderation in live video streams
• Distributed content filtering for objectionable segments
• Low-latency strobe removal for photosensitive viewers

Innovation

• Extends Media Over QUIC for real-time moderation
• Filters objectionable segments without stopping stream
• Distributes content analysis tasks to clients