Decentralized Video Provenance at the Edge: Blockchain-Assisted Integrity for Security Camera Streams

Marco Vela-Koentarjo, Madhusudan Singh

Keywords: Blockchain, Deepfakes, Media Verification, Security, Cryptography, IPFS, Smart Contracts

Abstract: With the rapid proliferation of generative AI, deepfakes and manipulated media pose increasing risks to political, economic, and social trust. The combination of accelerating developments in artificial intelligence and a lack of digital literacy creates an environment where malicious fake content can thrive. This work proposes a blockchain-based system for tamper evident registration and decentralized storage of security video footage. Unlike full camera-capture provenance, which guarantees end-to-end traceability from sensor to storage, our approach focuses on detectable post-capture tampering, providing verifiable logging of footage while defining clear threat and trust boundaries. By leveraging the Binance Smart Chain (BSC) for cryptographic logging and IPFS for decentralized storage, the system creates a foundation to mitigate the impact of AI manipulation in security and compliance applications. We evaluated the system by developing a functional prototype and conducting stress tests to assess cost efficiency and latency. Our analysis demonstrates that while the system provides robust integrity guarantees, decentralized storage introduces the primary latency bottleneck, with IPFS uploads averaging 2.08 seconds per segment. Furthermore, cost analysis indicates a baseline transaction fee of approximately $0.57 USD, which is viable for high-value evidence but requires Layer 2 scaling strategies for continuous streaming. Ultimately, this paper demonstrates that tamper-evident, blockchain-backed video registration is a technically feasible alternative to centralized custody logs, provided that scalability and storage considerations, such as decentralized pinning or permanent storage layers, are addressed.

Email Sharing: We authorize the sharing of all author emails with Program Chairs.

Data Release: We authorize the release of our submission and author names to the public in the event of acceptance.

Submission Number: 4