Asymmetric Phase Coding Video Watermarking

A training-free video watermark that embeds a complete Ed25519 signature into the chroma phase spectrum, enabling public verification without shared secrets.

Guang Yang; Fengchen Liu· 2026· DOI 10.48550/arXiv.2608.29212

The core problem

Existing video watermarking systems are symmetric: the party that can verify a mark holds the extractor weights or generator secret and can therefore also embed one. Benchmarks confirm the consequence, reporting that white-box forgery defeats all evaluated methods. This work presents a training-free video watermark that removes the shared secret. The signer embeds a complete Ed25519 signature into the phase spectrum of the chroma plane; any party holding the 32-byte public key and public per-video metadata verifies offline, with no model, no registry, and no network. The payload, 1024 bits of signed message with error correction, is an order of magnitude above common learned payloads.

On 1000 uncurated real-world clips, the system ships a verifying signature for 99.3% of the corpus and accepts a wrong public key zero times in 1000 attempts. An attack-aware acceptance gate yields embeddings that survive H.264 re-encoding at 100% and 50% rescaling at 97.4% on gated clips. The signature also verifies through a real display and capture loop, an axis absent from published evaluations. These results demonstrate high robustness and security, with a payload capacity of 1024 bits, an order of magnitude above common learned payloads.

Existing video watermarking systems are symmetric: the party that can verify a mark holds the extractor weights or generator secret and can therefore also embed one. Benchmarks confirm the consequence, reporting that white-box forgery defeats all evaluated methods. This work presents a training-free video watermark that removes the shared secret. The signer embeds a complete Ed25519 signature into the phase spectrum of the chroma plane; any party holding the 32-byte public key and public per-video metadata verifies offline, with no model, no registry, and no network. The payload, 1024 bits of signed message with error correction, is an order of magnitude above common learned payloads.

The proposed method relies on three design elements:

The asymmetric design eliminates the shared secret problem, enabling public verification without compromising security. The training-free approach avoids the need for large datasets or model training, making it practical for real-world deployment. The closed-loop signing procedure ensures that each video's embedding strength is optimized for reliable extraction, while the payload-free search provides resilience to geometric transformations. The system's ability to survive H.264 re-encoding and rescaling, as well as display-capture loops, highlights its potential for copyright protection and content authentication. Future work could explore extending the payload capacity and further improving robustness against adversarial attacks.

CS practitioners and researchers

Opening member content…