pith. sign in

arxiv: 2606.17809 · v1 · pith:UVH34MNCnew · submitted 2026-06-16 · 💻 cs.CV

Million-scale multimodal pollen microscopy with expert-guided foundation models

Andr\'as Biricz , Bj\"orn Gedda , Don\'at Magyar , Antonio Spanu , J\'anos Fillinger , P\'eter Pollner , Istv\'an Csabai This is my paper

Pith reviewed 2026-06-27 01:03 UTC · model grok-4.3

classification 💻 cs.CV
keywords pollen identificationmultimodal microscopyvision-language modelsdomain adaptationimage captioningaerobiologypalynology
0
0 comments X

The pith

Expert-guided vision-language models produce structured morphological captions for 1.5 million pollen grain detections across four regions and scanners.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper assembles Pollen AI Atlas, a resource of pure-species bright-field images from four geographic origins and four scanner settings covering 46 taxa. Token-level mining seeded by single exemplars yields 1,511,390 released detections at 99.6 percent proposal precision in test regions. Five open-weight VLMs generate grain-level captions describing aperture systems, wall ornamentation, shape and size under expert palynological anchors. Frozen visual features reach 88.16 percent top-1 accuracy on classification while caption-derived text embeddings maintain mAP@20 of 0.811 in cross-regional retrieval where image similarity falls to 0.262.

Core claim

A million-scale multimodal pollen microscopy resource is released in which each of 1,511,390 detections is paired with machine-generated structured captions; baseline benchmarks confirm that the resulting text embeddings remain effective for retrieval when visual domain shift degrades image similarity.

What carries the argument

Token-level mining and filtering seeded by one manually selected exemplar per slide, followed by expert-verified palynological anchors to guide five open-weight VLMs in producing structured morphological captions.

If this is right

  • The released data, annotations, captions, splits, code and weights form a public benchmark for pollen recognition.
  • Caption-derived embeddings support cross-regional domain adaptation where pure image similarity fails.
  • The pipeline demonstrates a route to domain-specific multimodal microscopy learning that retains palynological interpretability.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • Text embeddings generated under expert morphological anchors may generalise to other microscopy domains where visual domain shift is large but descriptive language remains stable.
  • The gap between image and text retrieval performance suggests that future work could test whether joint training on the paired image-caption data further closes the domain gap.
  • Releasing both the detections and the five-model caption sets allows direct comparison of how different VLMs affect downstream retrieval robustness.

Load-bearing premise

Expert-verified palynological anchors are sufficient to stop hallucination or factual drift in the VLMs when the grains come from scanner settings and geographic origins not seen during caption generation.

What would settle it

Measure caption factual accuracy and cross-regional retrieval mAP on a held-out fifth geographic origin or fifth scanner setting whose images were never used to create or verify the anchors.

Figures

Figures reproduced from arXiv: 2606.17809 by Andr\'as Biricz, Antonio Spanu, Bj\"orn Gedda, Don\'at Magyar, Istv\'an Csabai, J\'anos Fillinger, P\'eter Pollner.

Figure 1
Figure 1. Figure 1: The released captioned pollen-grain dataset is evaluated through complementary corpus-level, caption-level and [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: One-shot initialisation workflow on whole-slide data. For each slide, Stage 1 uses a manually selected pollen-grain [PITH_FULL_IMAGE:figures/full_fig_p011_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Token-level mining workflow for whole-slide pollen discovery. Stage 1 constructs a token-space query representation [PITH_FULL_IMAGE:figures/full_fig_p012_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Quality-filtering workflow applied to raw mining detections. Stage 1 re-embeds each query–candidate pair by query– [PITH_FULL_IMAGE:figures/full_fig_p015_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: VLM captioning workflow. Stage 1 prepares slide-specific captioning context from pure-species slides and [PITH_FULL_IMAGE:figures/full_fig_p016_5.png] view at source ↗
read the original abstract

Automated pollen identification from microscopy remains a bottleneck in aerobiology, palaeoecology and biodiversity monitoring, because scalable systems must generalise across specimen preparation, scanner settings and geographic origins while retaining palynological interpretability. To address this gap, we present a million-scale multimodal pollen microscopy resource, Pollen AI Atlas, assembled from pure-species whole-slide bright-field images spanning four geographic origins, four scanner settings and 46 taxon labels across 31 botanical families. Seeded by one manually selected exemplar per source slide, token-level mining and filtering produced 1,511,390 released grain detections with 99.6\% proposal precision in expert-curated test regions. Each detection was paired with machine-generated grain-level morphological captions from five open-weight vision-language models, guided by expert-verified palynological anchors, yielding structured descriptions of aperture systems, wall ornamentation, shape and size. Among the evaluated models, Gemma4 provided the most controlled primary caption set, combining tight length control, no leakage and the strongest text-retrieval performance. Baseline benchmarks with frozen visual features reached 88.16\% top-1 accuracy, while cross-regional retrieval showed that caption-derived text embeddings remained robust when image similarity degraded (mAP@20 0.811 versus 0.262). Released data, annotations, captions, splits, code, and weights provide a benchmark for pollen recognition, cross-regional domain adaptation and domain-specific multimodal microscopy learning.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

1 major / 1 minor

Summary. The paper presents Pollen AI Atlas, a million-scale multimodal pollen microscopy resource assembled from whole-slide bright-field images across four geographic origins, four scanner settings, and 46 taxa. Token-level mining seeded by one manual exemplar per slide yields 1,511,390 released detections at 99.6% proposal precision in expert-curated test regions; each detection is paired with structured morphological captions (aperture systems, wall ornamentation, shape, size) generated by five open-weight VLMs guided by expert-verified palynological anchors. Baseline frozen-feature benchmarks reach 88.16% top-1 accuracy, and caption-derived text embeddings show robust cross-regional retrieval (mAP@20 0.811 vs. 0.262 image similarity). The work releases data, annotations, captions, splits, code, and weights as a benchmark for pollen recognition and domain-specific multimodal learning.

Significance. If the captions prove reliable under domain shift, the resource supplies a large, structured, publicly released multimodal benchmark that directly supports research on cross-regional domain adaptation and palynological interpretability in microscopy. The purely empirical construction (no fitted parameters or invented entities) and release of code/weights are concrete strengths that increase the dataset's immediate utility for reproducible benchmarking.

major comments (1)
  1. [Abstract] Abstract: the central claim that the released captions provide reliable 'structured descriptions of aperture systems, wall ornamentation, shape and size' rests on seeding five open-weight VLMs with expert-verified palynological anchors, yet no quantitative factual-accuracy audit (e.g., expert review rates or error rates on caption fields) is reported for grains from scanner settings or geographic origins held out from the anchor set. This directly affects the validity of the multimodal benchmark and the reported mAP@20 0.811 for caption-derived embeddings.
minor comments (1)
  1. [Abstract] Abstract and methods: the exact filtering rules that produced the final 1,511,390 released detections are referenced but not enumerated; adding a concise list or pseudocode would improve reproducibility.

Simulated Author's Rebuttal

1 responses · 0 unresolved

We thank the referee for the detailed review and constructive comment on caption reliability. We address the concern point-by-point below.

read point-by-point responses

  1. Referee: [Abstract] Abstract: the central claim that the released captions provide reliable 'structured descriptions of aperture systems, wall ornamentation, shape and size' rests on seeding five open-weight VLMs with expert-verified palynological anchors, yet no quantitative factual-accuracy audit (e.g., expert review rates or error rates on caption fields) is reported for grains from scanner settings or geographic origins held out from the anchor set. This directly affects the validity of the multimodal benchmark and the reported mAP@20 0.811 for caption-derived embeddings.

    Authors: We agree that no quantitative factual-accuracy audit (expert review rates or per-field error rates on held-out scanner settings or geographic origins) is reported in the current manuscript. The captions rely on expert-verified palynological anchors to guide the VLMs, and the reported mAP@20 of 0.811 for text embeddings provides indirect support via retrieval performance. To directly address the referee's concern about reliability under domain shift, the revised manuscript will include a new expert audit: a random sample of captions from held-out regions will be reviewed by a palynologist, with error rates reported per morphological field (aperture systems, wall ornamentation, shape, size). This evaluation will be added to the Methods and Results sections. revision: yes

Circularity Check

0 steps flagged

No circularity; empirical data release with measured metrics

full rationale

The paper assembles and releases a multimodal pollen microscopy dataset via token-level mining and VLM captioning seeded by expert anchors. Reported figures (1,511,390 detections, 99.6% proposal precision, 88.16% top-1 accuracy, mAP@20 values) are direct empirical measurements from expert-curated test regions and baseline benchmarks, not quantities defined in terms of other internal quantities or reduced by construction. No equations, fitted parameters presented as predictions, self-citation load-bearing premises, or ansatzes appear in the abstract or described pipeline. The work is self-contained against external benchmarks as a data resource.

Axiom & Free-Parameter Ledger

0 free parameters · 2 axioms · 0 invented entities

The work is a data-curation and benchmarking effort that relies on standard computer-vision pipelines and off-the-shelf VLMs; no new physical or mathematical entities are postulated.

axioms (2)
  • domain assumption Pure-species whole-slide images contain grains belonging to a single botanical taxon.
    Invoked when seeding detections from one manually selected exemplar per source slide and when assigning the 46 taxon labels.
  • domain assumption Expert-verified palynological anchors are factually correct and sufficient to constrain VLM output.
    Central to the claim that the generated captions are structured and reliable descriptions of aperture systems, wall ornamentation, shape and size.

pith-pipeline@v0.9.1-grok · 5816 in / 1530 out tokens · 34275 ms · 2026-06-27T01:03:12.967794+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

74 extracted references · 40 canonical work pages · 1 internal anchor

  1. [1]

    OpenSlide: A vendor-neutral software foundation for digital pathology , journal =

    Adam Goode and Benjamin Gilbert and Jan Harkes and Drazen Jukic and Mahadev Satyanarayanan , keywords =. OpenSlide: A vendor-neutral software foundation for digital pathology , journal =. 2013 , issn =. doi:10.4103/2153-3539.119005 , url =

    work page doi:10.4103/2153-3539.119005 2013

  2. [2]

    PLOS ONE , publisher =

    Feature Extraction and Machine Learning for the Classification of Brazilian Savannah Pollen Grains , year =. PLOS ONE , publisher =. doi:10.1371/journal.pone.0157044 , author =

    work page doi:10.1371/journal.pone.0157044

  3. [3]

    PLOS ONE , publisher =

    Improving classification of pollen grain images of the POLEN23E dataset through three different applications of deep learning convolutional neural networks , year =. PLOS ONE , publisher =. doi:10.1371/journal.pone.0201807 , author =

    work page doi:10.1371/journal.pone.0201807

  4. [4]

    PLOS ONE , publisher =

    Precise automatic classification of 46 different pollen types with convolutional neural networks , year =. PLOS ONE , publisher =. doi:10.1371/journal.pone.0229751 , author =

    work page doi:10.1371/journal.pone.0229751

  5. [5]

    arXiv preprint arXiv:1802.03426 , year=

    UMAP: Uniform Manifold Approximation and Projection for Dimension Reduction , author=. arXiv preprint arXiv:1802.03426 , year=. 1802.03426 , archivePrefix=

    Pith/arXiv arXiv

  6. [6]

    ArXiv , year=

    Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks , author=. ArXiv , year=

  7. [7]

    ArXiv , year=

    DINOv2: Learning Robust Visual Features without Supervision , author=. ArXiv , year=

  8. [8]

    Efficient memory management for large language model serving with pagedattention

    Kwon, Woosuk and Li, Zhuohan and Zhuang, Siyuan and Sheng, Ying and Zheng, Lianmin and Yu, Cody Hao and Gonzalez, Joseph and Zhang, Hao and Stoica, Ion , title =. Proceedings of the 29th Symposium on Operating Systems Principles , pages =. 2023 , isbn =. doi:10.1145/3600006.3613165 , abstract =

    work page doi:10.1145/3600006.3613165 2023

  9. [9]

    2025 , eprint=

    Qwen2.5 Technical Report , author=. 2025 , eprint=

    2025

  10. [10]

    2025 , eprint=

    Qwen2.5-VL Technical Report , author=. 2025 , eprint=

    2025

  11. [11]

    2025 , eprint=

    Qwen3-VL Technical Report , author=. 2025 , eprint=

    2025

  12. [12]

    ArXiv , year=

    SAM 2: Segment Anything in Images and Videos , author=. ArXiv , year=

  13. [13]

    7th International Conference on Learning Representations, ICLR 2019, New Orleans, LA, USA, May 6-9, 2019 , publisher =

    Decoupled Weight Decay Regularization , author =. 7th International Conference on Learning Representations, ICLR 2019, New Orleans, LA, USA, May 6-9, 2019 , publisher =. 2019 , url =

    2019

  14. [14]

    and Cecchi, L

    D’Amato, G. and Cecchi, L. and Bonini, S. and Nunes, C. and Annesi-Maesano, I. and Behrendt, H. and Liccardi, G. and Popov, T. and Van Cauwenberge, P. , title =. Allergy , volume =. doi:10.1111/j.1398-9995.2007.01393.x , url =. https://onlinelibrary.wiley.com/doi/pdf/10.1111/j.1398-9995.2007.01393.x , abstract =

    work page doi:10.1111/j.1398-9995.2007.01393.x 2007

  15. [15]

    Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV) , month =

    Caron, Mathilde and Touvron, Hugo and Misra, Ishan and J\'egou, Herv\'e and Mairal, Julien and Bojanowski, Piotr and Joulin, Armand , title =. Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV) , month =. 2021 , pages =

    2021

  16. [16]

    and Bialic, \'E

    Cholleton, D. and Bialic, \'E. and Dumas, A. and Kaluzny, P. and Rairoux, P. and Miffre, A. , TITLE =. Atmospheric Measurement Techniques , VOLUME =. 2022 , NUMBER =

    2022

  17. [17]

    The EUMETNET AutoPollen programme: establishing a prototype automatic pollen monitoring network in Europe , journal=

    Clot, Bernard and Gilge, Stefan and Hajkova, Lenka and Magyar, Don. The EUMETNET AutoPollen programme: establishing a prototype automatic pollen monitoring network in Europe , journal=. 2024 , month=. doi:10.1007/s10453-020-09666-4 , url=

    work page doi:10.1007/s10453-020-09666-4 2024

  18. [18]

    Molecular Ecology Resources , volume =

    Lang, Dandan and Tang, Min and Hu, Jiahui and Zhou, Xin , title =. Molecular Ecology Resources , volume =. doi:10.1111/1755-0998.13061 , url =. https://onlinelibrary.wiley.com/doi/pdf/10.1111/1755-0998.13061 , abstract =

    work page doi:10.1111/1755-0998.13061

  19. [19]

    , booktitle=

    Dhamija, Akshay Raj and Günther, Manuel and Ventura, Jonathan and Boult, Terrance E. , booktitle=. The Overlooked Elephant of Object Detection: Open Set , year=

  20. [20]

    HIRST, J. M. , title =. Annals of Applied Biology , volume =. doi:10.1111/j.1744-7348.1952.tb00904.x , url =. https://onlinelibrary.wiley.com/doi/pdf/10.1111/j.1744-7348.1952.tb00904.x , abstract =

    work page doi:10.1111/j.1744-7348.1952.tb00904.x 1952

  21. [21]

    Holt and G

    K. Holt and G. Allen and R. Hodgson and S. Marsland and J. Flenley , keywords =. Progress towards an automated trainable pollen location and classifier system for use in the palynology laboratory , journal =. 2011 , issn =. doi:10.1016/j.revpalbo.2011.08.006 , url =

    work page doi:10.1016/j.revpalbo.2011.08.006 2011

  22. [22]

    2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) , year=

    Towards Open World Object Detection , author=. 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) , year=

    2021

  23. [23]

    Lozano, Alejandro and Sun, Min Woo and Burgess, James and Chen, Liangyu and Nirschl, Jeffrey J. and Gu, Jeffrey and Lopez, Ivan and Aklilu, Josiah and Rau, Anita and Katzer, Austin Wolfgang and Zhang, Yuhui and Chiu, Collin and Wang, Xiaohan and Song, Alfred Seunghoon and Tibshirani, Robert and Yeung-Levy, Serena , title =. Proceedings of the IEEE/CVF Con...

    2025

  24. [24]

    Macenko, Marc and Niethammer, Marc and Marron, J. S. and Borland, David and Woosley, John T. and Xiaojun Guan and Schmitt, Charles and Thomas, Nancy E. , booktitle=. A method for normalizing histology slides for quantitative analysis , year=

  25. [25]

    Detection and Classification of Pollen Grain Microscope Images , year=

    Battiato, Sebastiano and Ortis, Alessandro and Trenta, Francesca and Ascari, Lorenzo and Politi, Mara and Siniscalco, Consolata , booktitle=. Detection and Classification of Pollen Grain Microscope Images , year=

  26. [26]

    Pollen monitoring: minimum requirements and reproducibility of analysis , journal=

    Gal. Pollen monitoring: minimum requirements and reproducibility of analysis , journal=. 2014 , month=. doi:10.1007/s10453-014-9335-5 , url=

    work page doi:10.1007/s10453-014-9335-5 2014

  27. [27]

    Micro-Bench: A Microscopy Benchmark for Vision-Language Understanding , url =

    Lozano, Alejandro and Nirschl, Jeffrey and Burgess, James and Gupte, Sanket Rajan and Zhang, Yuhui and Unell, Alyssa and Yeung-Levy, Serena , booktitle =. Micro-Bench: A Microscopy Benchmark for Vision-Language Understanding , url =. doi:10.52202/079017-0965 , editor =

    work page doi:10.52202/079017-0965

  28. [28]

    Pattern recognition methodologies for pollen grain image classification: a survey , journal=

    Viertel, Philipp and K. Pattern recognition methodologies for pollen grain image classification: a survey , journal=. 2022 , month=. doi:10.1007/s00138-021-01271-w , url=

    work page doi:10.1007/s00138-021-01271-w 2022

  29. [29]

    Automated recognition by multiple convolutional neural networks of modern, fossil, intact and damaged pollen grains , journal =

    Benjamin Bourel and Ross Marchant and Thibault. Automated recognition by multiple convolutional neural networks of modern, fossil, intact and damaged pollen grains , journal =. 2020 , issn =. doi:10.1016/j.cageo.2020.104498 , url =

    work page doi:10.1016/j.cageo.2020.104498 2020

  30. [30]

    Illustrated Pollen Terminology , publisher =

    Halbritter, Heidemarie and Silvia, Ulrich and Grímsson, Friðgeir and Weber, Martina and Zetter, Reinhard and Hesse, Michael and Buchner, Ralf and Svojtka, Matthias and Frosch-Radivo, Andrea , year =. Illustrated Pollen Terminology , publisher =

  31. [31]

    Holt, K. A. and Bennett, K. D. , title =. New Phytologist , volume =. doi:10.1111/nph.12848 , url =. https://nph.onlinelibrary.wiley.com/doi/pdf/10.1111/nph.12848 , abstract =

    work page doi:10.1111/nph.12848

  32. [32]

    2026 , url =

    Qwen Team , month =. 2026 , url =

    2026

  33. [33]

    , title=

    Ravi, Jampani and Narmadha, R. , title=. Scientific Reports , year=. doi:10.1038/s41598-024-81594-6 , url=

    work page doi:10.1038/s41598-024-81594-6

  34. [34]

    POLLEN73S: An image dataset for pollen grains classification , journal =

    Gilberto Astolfi and Ariadne Barbosa Gonçalves and Geazy Vilharva Menezes and Felipe Silveira Brito Borges and Angelica Christina Melo Nunes Astolfi and Edson Takashi Matsubara and Marco Alvarez and Hemerson Pistori , keywords =. POLLEN73S: An image dataset for pollen grains classification , journal =. 2020 , issn =. doi:10.1016/j.ecoinf.2020.101165 , url =

    work page doi:10.1016/j.ecoinf.2020.101165 2020

  35. [35]

    Assisting monofloral honey classification by automated pollen identification based on convolutional neural networks , journal =

    José Miguel Valiente and Juan José Martín-Osuna and Ana María Peral and Isabel Escriche , keywords =. Assisting monofloral honey classification by automated pollen identification based on convolutional neural networks , journal =. 2025 , issn =. doi:10.1016/j.ecoinf.2025.103340 , url =

    work page doi:10.1016/j.ecoinf.2025.103340 2025

  36. [36]

    A new learning paradigm: Learning using privileged information , journal =

    Vladimir Vapnik and Akshay Vashist , keywords =. A new learning paradigm: Learning using privileged information , journal =. 2009 , note =. doi:10.1016/j.neunet.2009.06.042 , url =

    work page doi:10.1016/j.neunet.2009.06.042 2009

  37. [37]

    Method development and application of object detection and classification to Quaternary fossil pollen sequences , journal =

    Robin. Method development and application of object detection and classification to Quaternary fossil pollen sequences , journal =. 2024 , issn =. doi:10.1016/j.quascirev.2024.108521 , url =

    work page doi:10.1016/j.quascirev.2024.108521 2024

  38. [38]

    International Conference on Learning Representations , journal =

    Alexey Dosovitskiy and Lucas Beyer and Alexander Kolesnikov and Dirk Weissenborn and Xiaohua Zhai and Thomas Unterthiner and Mostafa Dehghani and Matthias Minderer and Georg Heigold and Sylvain Gelly and Jakob Uszkoreit and Neil Houlsby , title =. International Conference on Learning Representations , journal =. 2020 , url =

    2020

  39. [39]

    Deep learning of pollen images under low annotation costs: joint optimization of morphological features and training and prediction strategies , journal =

    Teng Zhang and Limi Mao , keywords =. Deep learning of pollen images under low annotation costs: joint optimization of morphological features and training and prediction strategies , journal =. 2026 , issn =. doi:10.1016/j.revpalbo.2025.105458 , url =

    work page doi:10.1016/j.revpalbo.2025.105458 2026

  40. [40]

    The Visual Computer , year=

    Zu, Baokai and Cao, Tong and Li, Yafang and Li, Jianqiang and Wang, Hongyuan and Wang, Quanzeng , title=. The Visual Computer , year=. doi:10.1007/s00371-024-03701-y , url=

    work page doi:10.1007/s00371-024-03701-y

  41. [41]

    On the Effectiveness of ViT Features as Local Semantic Descriptors

    Amir, Shir and Gandelsman, Yossi and Bagon, Shai and Dekel, Tali. On the Effectiveness of ViT Features as Local Semantic Descriptors. Computer Vision -- ECCV 2022 Workshops. 2023

    2022

  42. [42]

    Leitfaden der

    Beug, Hans-J. Leitfaden der. 2021 , pages =

    2021

  43. [43]

    SAM 3: Segment Anything with Concepts

    SAM 3: Segment Anything with Concepts , author=. 2026 , eprint=. doi:10.48550/arXiv.2511.16719 , note=

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.2511.16719 2026

  44. [44]

    Jardine, P. E. and Gosling, W. D. and Lomax, B. H. and Julier, A. C. M. and Fraser, W. T. , TITLE =. Journal of Micropalaeontology , VOLUME =. 2019 , NUMBER =

    2019

  45. [45]

    , title=

    Barnes, Charles S. , title=. Current Allergy and Asthma Reports , year=. doi:10.1007/s11882-018-0813-7 , url=

    work page doi:10.1007/s11882-018-0813-7

  46. [46]

    Proceedings of the 38th International Conference on Machine Learning , pages =

    Learning Transferable Visual Models From Natural Language Supervision , author =. Proceedings of the 38th International Conference on Machine Learning , pages =. 2021 , editor =

    2021

  47. [47]

    , booktitle=

    Yang, Nana and Joos, Victor and Jacquemart, A -L and Buyens, Christel and De Vleeschouwer, C. , booktitle=. Using Pure Pollen Species When Training a CNN to Segment Pollen Mixtures , year=

  48. [48]

    and Chen, Bowen and Williamson, Drew F

    Lu, Ming Y. and Chen, Bowen and Williamson, Drew F. K. and Chen, Richard J. and Liang, Ivy and Ding, Tong and Jaume, Guillaume and Odintsov, Igor and Le, Long Phi and Gerber, Georg and Parwani, Anil V. and Zhang, Andrew and Mahmood, Faisal , title=. Nature Medicine , year=. doi:10.1038/s41591-024-02856-4 , url=

    work page doi:10.1038/s41591-024-02856-4

  49. [49]

    , title =

    Dunker, Susanne and Motivans, Elena and Rakosy, Demetra and Boho, David and Mäder, Patrick and Hornick, Thomas and Knight, Tiffany M. , title =. New Phytologist , volume =. doi:10.1111/nph.16882 , url =. https://nph.onlinelibrary.wiley.com/doi/pdf/10.1111/nph.16882 , abstract =

    work page doi:10.1111/nph.16882

  50. [50]

    , year =

    Selesnick, Ivan and Baraniuk, Richard and Kingsbury, N.C. , year =. The dual-tree complex wavelet transform , volume =. Signal Processing Magazine, IEEE , doi =

  51. [51]

    Evaluation of Pollen Apps Forecasts: The Need for Quality Control in an eHealth Service

    Bastl, Katharina and Berger, Uwe and Kmenta, Maximilian. Evaluation of Pollen Apps Forecasts: The Need for Quality Control in an eHealth Service. J Med Internet Res. 2017

    2017

  52. [52]

    The allergy risk associated with pollens in France [Risque allergique lié aux pollens en France] , volume =

    Thibaudon, Michel , year =. The allergy risk associated with pollens in France [Risque allergique lié aux pollens en France] , volume =. European Annals of Allergy and Clinical Immunology , abstract =

  53. [53]

    PLOS ONE , publisher =

    Pollen identification through convolutional neural networks: First application on a full fossil pollen sequence , year =. PLOS ONE , publisher =. doi:10.1371/journal.pone.0302424 , author =

    work page doi:10.1371/journal.pone.0302424

  54. [54]

    2025 , url =

    Cupitt, John and Martinez, Kleis , title =. 2025 , url =

    2025

  55. [55]

    and Belmonte, Jordina and De Linares, Concepci

    Polling, Marcel and Li, Chen and Cao, Lu and Verbeek, Fons and de Weger, Letty A. and Belmonte, Jordina and De Linares, Concepci. Neural networks for increased accuracy of allergenic pollen monitoring , journal=. 2021 , month=. doi:10.1038/s41598-021-90433-x , url=

    work page doi:10.1038/s41598-021-90433-x 2021

  56. [56]

    Computer Vision – ECCV 2022: 17th European Conference, Tel Aviv, Israel, October 23–27, 2022, Proceedings, Part X , pages =

    Minderer, Matthias and Gritsenko, Alexey and Stone, Austin and Neumann, Maxim and Weissenborn, Dirk and Dosovitskiy, Alexey and Mahendran, Aravindh and Arnab, Anurag and Dehghani, Mostafa and Shen, Zhuoran and Wang, Xiao and Zhai, Xiaohua and Kipf, Thomas and Houlsby, Neil , title =. Computer Vision – ECCV 2022: 17th European Conference, Tel Aviv, Israel,...

    work page doi:10.1007/978-3-031-20080-9_42 2022

  57. [57]

    How to identify pollen like a palynologist: A prior knowledge-guided deep feature learning for real-world pollen classification , journal =

    Jianqiang Li and Wenxiu Cheng and Xi Xu and Linna Zhao and Suqin Liu and Zhengkai Gao and Caihua Ye and Huanling You , keywords =. How to identify pollen like a palynologist: A prior knowledge-guided deep feature learning for real-world pollen classification , journal =. 2024 , issn =. doi:10.1016/j.eswa.2023.121392 , url =

    work page doi:10.1016/j.eswa.2023.121392 2024

  58. [58]

    Nature634(8033), 466–473 (Oct 2024)

    Lu, Ming Y. and Chen, Bowen and Williamson, Drew F. K. and Chen, Richard J. and Zhao, Melissa and Chow, Aaron K. and Ikemura, Kenji and Kim, Ahrong and Pouli, Dimitra and Patel, Ankush and Soliman, Amr and Chen, Chengkuan and Ding, Tong and Wang, Judy J. and Gerber, Georg and Liang, Ivy and Le, Long Phi and Parwani, Anil V. and Weishaupt, Luca L. and Mahm...

    work page doi:10.1038/s41586-024-07618-3

  59. [59]

    POLLEN13K: A Large Scale Microscope Pollen Grain Image Dataset , year=

    Battiato, Sebastiano and Ortis, Alessandro and Trenta, Francesca and Ascari, Lorenzo and Politi, Mara and Siniscalco, Consolata , booktitle=. POLLEN13K: A Large Scale Microscope Pollen Grain Image Dataset , year=

  60. [60]

    and Ammann, Lolita and Entling, Martin H

    Bertrand, Colette and Eckerter, Philipp W. and Ammann, Lolita and Entling, Martin H. and Gobet, Erika and Herzog, Felix and Mestre, Laia and Tinner, Willy and Albrecht, Matthias , title =. Journal of Applied Ecology , volume =. doi:10.1111/1365-2664.13483 , url =. https://besjournals.onlinelibrary.wiley.com/doi/pdf/10.1111/1365-2664.13483 , abstract =

    work page doi:10.1111/1365-2664.13483

  61. [61]

    and Smith, Henrik G

    Olsson, Ola and Karlsson, Melanie and Persson, Anna S. and Smith, Henrik G. and Varadarajan, Vidula and Yourstone, Johanna and Stjernman, Martin , title =. Methods in Ecology and Evolution , volume =. doi:10.1111/2041-210X.13575 , url =. https://besjournals.onlinelibrary.wiley.com/doi/pdf/10.1111/2041-210X.13575 , abstract =

    work page doi:10.1111/2041-210x.13575 2041

  62. [62]

    Shariar Hossain and Ahamed, Md

    Islam, Syeda Munjiba and Sazzad, Md. Shariar Hossain and Ahamed, Md. Faysal and Ahmed, Maaz and Nayon, Sadikul Islam and Islam, Toufiq and Ayon, Sabrina Saniat and Haider, Julfikar and Abdullah-Al-Wadud, M. and Islam, S. M. Riazul , title=. Scientific Data , year=. doi:10.1038/s41597-025-06264-2 , url=

    work page doi:10.1038/s41597-025-06264-2

  63. [63]

    Scientific Reports , doi =

    Bankhead, Peter and Loughrey, Maurice and Fernandez, Jose and Dombrowski, Yvonne and Mcart, Darragh and Dunne, Philip and Mcquaid, Stephen and Gray, Ronan and Murray, Liam and Coleman, Helen and James, Jacqueline and Salto-Tellez, Manuel and Hamilton, Peter , year =. Scientific Reports , doi =

  64. [64]

    Efficient and scalable training set generation for automated pollen monitoring with

    Biricz, Andr. Efficient and scalable training set generation for automated pollen monitoring with. Scientific Reports , year=. doi:10.1038/s41598-025-31646-2 , url=

    work page doi:10.1038/s41598-025-31646-2

  65. [65]

    The FAIR Guiding Principles for scientific data management and stewardship , volume =

    Wilkinson, Mark and Dumontier, Michel and Aalbersberg, IJsbrand Jan and Appleton, Gaby and Axton, Myles and Baak, Arie and Blomberg, Niklas and Boiten, Jan-Willem and Bonino da Silva Santos, Luiz Olavo and Bourne, Philip and Bouwman, Jildau and Brookes, Anthony and Clark, Tim and Crosas, Merce and Dillo, Ingrid and Dumon, Olivier and Edmunds, Scott and Ev...

  66. [66]

    2025 , eprint=

    DINOv3 , author=. 2025 , eprint=

    2025

  67. [67]

    Pollen Season Trends (1973-2013) in Stockholm Area, Sweden , volume =

    Lind, Tomas and Ekebom, Agneta and Kübler, Kerstin and Östensson, Pia and Bellander, Tom and Lõhmus, Mare , year =. Pollen Season Trends (1973-2013) in Stockholm Area, Sweden , volume =. PLOS ONE , doi =

    1973

  68. [68]

    and Zeder, Y

    Sauvageat, E. and Zeder, Y. and Auderset, K. and Calpini, B. and Clot, B. and Crouzy, B. and Konzelmann, T. and Lieberherr, G. and Tummon, F. and Vasilatou, K. , TITLE =. Atmospheric Measurement Techniques , VOLUME =. 2020 , NUMBER =

    2020

  69. [69]

    Recommended terminology for aerobiological studies: automatic and real-time monitoring methods , volume =

    Tummon, Fiona and Bastl, Katharina and Bastl, Maximilian and Beggs, Paul and Belmonte, Jordina and Bruffaerts, Nicolas and Charalampopoulos, Athanasios and Clot, Bernard and Cristofori, Antonella and Crouzy, Benoît and Damialis, Athanasios and Erb, Sophie and Galán, Carmen and Graf, Elias and Haus, Jörg and Koch, Kilian and Lieberherr, Gian and Lbadaoui-D...

  70. [70]

    doi:10.5281/zenodo.6327079 , url =

    Poehlmann, Andreas and Villalba, Santiago , title =. doi:10.5281/zenodo.6327079 , url =

    work page doi:10.5281/zenodo.6327079

  71. [71]

    and Song, Andrew H

    Ding, Tong and Wagner, Sophia J. and Song, Andrew H. and Chen, Richard J. and Lu, Ming Y. and Zhang, Andrew and Vaidya, Anurag J. and Jaume, Guillaume and Shaban, Muhammad and Kim, Ahrong and Williamson, Drew F. K. and Robertson, Harry and Chen, Bowen and Almagro-P. A multimodal whole-slide foundation model for pathology , journal=. 2025 , month=. doi:10....

    work page doi:10.1038/s41591-025-03982-3 2025

  72. [72]

    Pollen recognition through an open-source web-based system: automated particle counting for aerobiological analysis , journal=

    Chaves, Antonio Jes. Pollen recognition through an open-source web-based system: automated particle counting for aerobiological analysis , journal=. 2024 , month=. doi:10.1007/s12145-023-01189-z , url=

    work page doi:10.1007/s12145-023-01189-z 2024

  73. [73]

    Automated multifocus pollen detection using deep learning , journal=

    Gallardo, Ram. Automated multifocus pollen detection using deep learning , journal=. 2024 , month=. doi:10.1007/s11042-024-18450-2 , url=

    work page doi:10.1007/s11042-024-18450-2 2024

  74. [74]

    2026 , publisher =

    Biricz, András and Gedda, Björn and Magyar, Donát and Spanu, Antonio and Fillinger, János and Pollner, Péter and Csabai, István , title =. 2026 , publisher =. doi:10.5281/zenodo.20690944 , url =

    work page doi:10.5281/zenodo.20690944 2026