Metraipolis › archive › dc_afa8eb9ad586
Computationally Efficient Pathology Segmentation using Knowledge Distillation from Foundation Models — arXiv:2609.03947
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2609.03947
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-17T03:23:46.438Z
- SHA-256 of the archived text
3c4043eb57604d95dbb0207c0af37b5b7016a786c5419f1626882a01f10833ea- Stable URL of this page
https://metraipolis.com/documento/dc_afa8eb9ad586
How to cite
Computationally Efficient Pathology Segmentation using Knowledge Distillation from Foundation Models — arXiv:2609.03947 — arXiv (en), https://arxiv.org/abs/2609.03947 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_afa8eb9ad586 (Metraipolis archive dc_afa8eb9ad586, sha256 3c4043eb57604d95dbb0207c0af37b5b7016a786c5419f1626882a01f10833ea)
BibTeX
@misc{dcafa8eb9ad586,
title = {Computationally Efficient Pathology Segmentation using Knowledge Distillation from Foundation Models — arXiv:2609.03947},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2609.03947},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_afa8eb9ad586; sha256 3c4043eb57604d95dbb0207c0af37b5b7016a786c5419f1626882a01f10833ea},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.03947. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Automatic tissue segmentation is essential for large-scale analysis of histopathology whole-slide images (WSIs), but accurate pixel-level segmentation remains challenging. Pixel-level annotations are expensive to obtain, models pre-trained on natural images may transfer poorly to histopathology, and pathology foundation models, despite their strong representations, are computationally expensive to deploy at scale. We address these challenges with a foundation-model knowledge distillation framework for efficient tissue segmentation. We first train Virchow2-based segmentation teachers, including a LoRA-adapted variant, achieving state-of-the-art or highly competitive performance across four datasets: PUMA, IGNITE, BEETLE, and a private blood vessel segmentation dataset. We then transfer response-level and feature-level knowledge from these teachers into compact student networks. Distillation consistently improves student performance over supervised training alone, producing state-of-the-art or near state-of-the-art results with substantially fewer parameters and up to ten-fold higher inference throughput than foundation-model-based segmentation networks. These results show that foundation-model knowledge can be effectively transferred to efficient segmentation models for scalable deployment without compromising performance. Models trained on the public datasets will be released through TIAToolbox. Tema degli esploratori: statistica e scienza dei dati (retro-catalogo arXiv: annunci del 2026-09-04 in eess.IV). Categorie arXiv: Image and Video Processing (eess.IV). Pagina del lavoro: https://arxiv.org/abs/2609.03947 — 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 2026-10-17T03:12:36.748Z dalla pagina https://arxiv.org/catchup/eess.IV/2026-09-04?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).