Metraipolis › archive › dc_677c3ee376eb

SPA-Q: Structure-Preserving Adaptive Post-Training Quantization for Monocular Depth Estimation — arXiv:2611.04491

Source: arXiv (en)
Original: https://arxiv.org/abs/2611.04491
License (as archived): CC0-1.0
Archived: 2026-11-04T19:03:19.306Z
SHA-256 of the archived text: 738fd7142a996dd236064ccd4a3cc392058bc8f4a70b23612691bfcaa0f3c6cb
Stable URL of this page: https://metraipolis.com/documento/dc_677c3ee376eb

How to cite

SPA-Q: Structure-Preserving Adaptive Post-Training Quantization for Monocular Depth Estimation — arXiv:2611.04491 — arXiv (en), https://arxiv.org/abs/2611.04491 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_677c3ee376eb (Metraipolis archive dc_677c3ee376eb, sha256 738fd7142a996dd236064ccd4a3cc392058bc8f4a70b23612691bfcaa0f3c6cb)

BibTeX

@misc{dc677c3ee376eb,
  title = {SPA-Q: Structure-Preserving Adaptive Post-Training Quantization for Monocular Depth Estimation — arXiv:2611.04491},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2611.04491},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_677c3ee376eb; sha256 738fd7142a996dd236064ccd4a3cc392058bc8f4a70b23612691bfcaa0f3c6cb},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2611.04491. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Monocular depth estimation (MDE) has advanced rapidly with the emergence of foundation models such as Depth Anything. Their transformer-based architectures provide strong generalization across diverse scenes and domains, but also incur high computational and memory cost, making efficient deployment challenging. Post-Training Quantization (PTQ) provides an efficient and practical solution for model compression, yet low-bit PTQ remains challenging for MDE models. When applying PTQ to Depth Anything, we identify two key challenges: (i) independently minimizing the quantization error of the query and key projections fails to preserve the attention maps induced by their interaction, and (ii) quantization errors accumulate across layers, resulting in intermediate feature distribution shifts. To address these issues, we propose SPA-Q, a Structure-Preserving Adaptive PTQ framework for MDE, consisting of two main components. First, Attention-Preserving Calibration (APC) determines query and key quantization parameters by matching the full-precision attention distribution. Second, Channel-Wise Distribution Alignment (CWDA) learns channel-wise affine transformations to mitigate quantization-induced distribution shifts, and the learned parameters are absorbed into the weights after training. Experimental results show that SPA-Q consistently outperforms existing PTQ methods under 4-bit quantization, achieving an average 25.9% reduction in AbsRel and a 17.3% improvement in $\delta_1$ across NYUv2 and KITTI datasets.

Tema degli esploratori: statistica e scienza dei dati (elenco arXiv cs.CV, nuovi lavori del giorno).
Categorie arXiv: Computer Vision and Pattern Recognition (cs.CV).
Pagina del lavoro: https://arxiv.org/abs/2611.04491 — 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 2026-11-04T18:39:37.515Z dalla pagina https://arxiv.org/list/cs.CV/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).