Paper Detail

StreamErase: Towards Ultra-Fast Streaming Video Erasing

Zhuobai Dong, Xiaobing Tu, Xiantao Zhang, Jinkui Ren, Yinggui Wang, Yue Ma, Linfeng Zhang

huggingface Score 4.0

Published 2026-10-14 · First seen 2026-10-15

General AI

Abstract

Current video object removal methods suffer from prohibitive latency (minutes per video) and non-causal processing that prevents real-time interaction. We present StreamErase, enabling sub-second latency with up to 26 FPS streaming generation on a single GPU. Our approach first augments a text-to-video model with mask input to generate high-quality videos that erase target objects and synthesize coherent backgrounds, but does not perform inference on the fly. As such, we progressively transfer the erasing capability from a powerful bidirectional foundation model to an efficient unidirectional streaming eraser, enabling long-horizon erasing without sacrificing visual fidelity. Several key challenges arise when generating videos over long, potentially infinite horizons: (1) bridging the domain gap from training on finite length and extrapolating to long horizons, (2) sustaining high quality by preventing error accumulation, and (3) maintaining fast inference, without increasing computational cost as the context grows. To address these challenges, we carefully designed sliding-window causal attention with attention sinks. By incorporating self-rollout with attention sinks and KV cache rolling during training, we simulate inference-time extrapolations with a fixed context window, enabling constant-speed generation of arbitrarily long videos. Finally, we establish a dedicated benchmark, LongErase, for streaming long-video object removal. Our model achieves strong performance compared to existing video object erasing models and generalizes well to real-world video scenarios, uniquely enabling infinite-length streaming. With StreamErase, users can input a mask and see the erasing results unfold in real-time, delivering a truly interactive experience.

Workflow Status

Review status
pending
Role
unreviewed
Read priority
later
Vote
Not set.
Saved
no
Collections
Not filed yet.
Next action
Not filled yet.

Reading Brief

No structured notes yet. Add `summary_sections`, `why_relevant`, `claim_impact`, or `next_action` in `papers.jsonl` to enrich this view.

Why It Surfaced

No ranking explanation is available yet.

Tags

No tags.

BibTeX

@misc{dong2026streamerase,
  title = {StreamErase: Towards Ultra-Fast Streaming Video Erasing},
  author = {Zhuobai Dong and Xiaobing Tu and Xiantao Zhang and Jinkui Ren and Yinggui Wang and Yue Ma and Linfeng Zhang},
  year = {2026},
  abstract = {Current video object removal methods suffer from prohibitive latency (minutes per video) and non-causal processing that prevents real-time interaction. We present StreamErase, enabling sub-second latency with up to 26 FPS streaming generation on a single GPU. Our approach first augments a text-to-video model with mask input to generate high-quality videos that erase target objects and synthesize coherent backgrounds, but does not perform inference on the fly. As such, we progressively transfer t},
  url = {https://huggingface.co/papers/2610.20137},
  keywords = {text-to-video model, bidirectional foundation model, unidirectional streaming eraser, sliding-window causal attention, attention sinks, KV cache rolling, self-rollout, streaming generation, code available, huggingface daily},
  eprint = {2610.20137},
  archiveprefix = {arXiv},
}

Metadata

{}