Metraipolis › archive › dc_54bea338eaee

Variable-Granularity Tokenization for High-Resolution Object Detection — arXiv:2608.28706

Source: arXiv (en)
Original: https://arxiv.org/abs/2608.28706
License (as archived): CC0-1.0
Archived: 2026-10-17T09:53:18.418Z
SHA-256 of the archived text: f364b7aeefc52cbdd0c6a899119198527bb7a470ec75fba8dea6a03bb1df39a5
Stable URL of this page: https://metraipolis.com/documento/dc_54bea338eaee

How to cite

Variable-Granularity Tokenization for High-Resolution Object Detection — arXiv:2608.28706 — arXiv (en), https://arxiv.org/abs/2608.28706 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_54bea338eaee (Metraipolis archive dc_54bea338eaee, sha256 f364b7aeefc52cbdd0c6a899119198527bb7a470ec75fba8dea6a03bb1df39a5)

BibTeX

@misc{dc54bea338eaee,
  title = {Variable-Granularity Tokenization for High-Resolution Object Detection — arXiv:2608.28706},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2608.28706},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_54bea338eaee; sha256 f364b7aeefc52cbdd0c6a899119198527bb7a470ec75fba8dea6a03bb1df39a5},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.28706. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

ViT detectors fix a uniform token grid before any learned stage. A native-resolution aerial detector must then choose between resolving few-pixel objects and staying inside compute and memory limits. We introduce VGTok, a training-free tokenizer that sets patch granularity per region from pixels, ahead of the encoder. VGTok scores each region by multi-scale morphological top-hat separability from its surround, then thresholds those scores at a per-image percentile, which fixes the token budget. A structure-tensor gate ($\lambda_{\min}$) refines only where two-dimensional object structure supports it, leaving one-dimensional clutter coarse. The resulting token set is a strict partition of the image. In a Co-DETR detector with an EVA-02 ViT-L encoder, VGTok clears every published VisDrone-val AP and AP$_S$ at every budget from 40\% to 100\% of tokens. At 40\% it records 44.22 AP with three fifths of the sequence discarded before the first transformer block; dense, it reaches 48.38 AP, $6.08$ above the strongest published entry. VGTok transfers to AI-TOD-v2 untouched, same scorer and same rank, and sets a new state of the art at 37.27 AP and 19.51 AP$_{vt}$. As a pure drop-in into a frozen checkpoint it reaches 36.29 AP at 78.5\% of tokens, above every published entry, where our 376.3M-parameter detector clears a 3.0B multi-expert model. We show that a token budget fixed before the backbone, from local separability and structure geometry alone, holds accuracy on the tiny-object regimes that dominate aerial detection, at $3.1\times$ less encoder compute and $1.9\times$ less encoder memory. Code and models are available at \href{ this https URL }{\texttt{ this http URL }} and \href{ this https URL }{\texttt{ this http URL }}.

Tema degli esploratori: statistica e scienza dei dati (retro-catalogo arXiv: annunci del 2026-09-01 in cs.CV).
Categorie arXiv: Computer Vision and Pattern Recognition (cs.CV).
Pagina del lavoro: https://arxiv.org/abs/2608.28706 — 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 2026-10-17T09:28:47.947Z dalla pagina https://arxiv.org/catchup/cs.CV/2026-09-01?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).