Metraipolis › archive › dc_ebc51ff5ce14

TV-AudioRemover: Joint Text-Visual Guided Sound Removal with Multi-Task Hard-Mixture Curriculum — arXiv:2609.25864

Source: arXiv (en)
Original: https://arxiv.org/abs/2609.25864
License (as archived): CC0-1.0
Archived: 2026-10-08T16:43:09.528Z
SHA-256 of the archived text: 848397bb582d3f21405126b26570a7c4bd4a3b995f2b2d127358e85c06357ea4
Stable URL of this page: https://metraipolis.com/documento/dc_ebc51ff5ce14

How to cite

TV-AudioRemover: Joint Text-Visual Guided Sound Removal with Multi-Task Hard-Mixture Curriculum — arXiv:2609.25864 — arXiv (en), https://arxiv.org/abs/2609.25864 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_ebc51ff5ce14 (Metraipolis archive dc_ebc51ff5ce14, sha256 848397bb582d3f21405126b26570a7c4bd4a3b995f2b2d127358e85c06357ea4)

BibTeX

@misc{dcebc51ff5ce14,
  title = {TV-AudioRemover: Joint Text-Visual Guided Sound Removal with Multi-Task Hard-Mixture Curriculum — arXiv:2609.25864},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.25864},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_ebc51ff5ce14; sha256 848397bb582d3f21405126b26570a7c4bd4a3b995f2b2d127358e85c06357ea4},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.25864. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Visual object removal can eliminate a target from video frames, yet its acoustic trace persists in the soundtrack, causing obvious audio-visual inconsistency. Existing video inpainting models operate solely on pixels, while audio editing models, especially for the sound removal task, are typically driven by text and therefore rely on limited single-modal control, which is less effective than multimodal guidance that provides stronger semantic grounding and temporal synchronization cues. In this paper, we present Text-Visual Guided Sound Removal (TV-AudioRemover), a target sound removal framework that leverages the visually edited video together with a natural-language instruction to suppress the sound associated with the removed visual object from the original audio mixture. To acquire high-quality training data, we devise a pipeline to construct a million-scale dataset of single-object audio-visual aligned samples, from which we synthesize mixture-target pairs customized for model training. To effectively leverage visual context and follow instruction intent, we augment the model architecture with task tokens, generalizable instruction modeling, and modality-specific global guidance. We further adopt multi-task training to strengthen task-role comprehension, and employ a hard-mixture curriculum that leverages semantically similar acoustic mixtures during fine-tuning to enhance fine-grained source discrimination. To support evaluation, we present AV-Remove-Bench, a comprehensive audio-visual object removal benchmark, along with dedicated objective metrics and an MLLM-based evaluation protocol. Experiments demonstrate that our method achieves state-of-the-art performance on both subjective and objective metrics. Project page: this https URL .

Tema degli esploratori: statistica e scienza dei dati (retro-catalogo arXiv: annunci del 2026-09-23 in cs.CV).
Categorie arXiv: Multimedia (cs.MM) ; Computer Vision and Pattern Recognition (cs.CV); Sound (cs.SD).
Pagina del lavoro: https://arxiv.org/abs/2609.25864 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-10-08T15:56:18.243Z dalla pagina https://arxiv.org/catchup/cs.CV/2026-09-23?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).