Image Dataset  ·  Aerobiology  ·  Computer Vision

POLLEN13K

A Large Scale Microscope Pollen Grain Image Dataset

Sebastiano Battiato  ·  Alessandro Ortis  ·  Francesca Trenta
University of Catania – Department of Mathematics and Computer Science

Lorenzo Ascari  ·  Mara Politi  ·  Consolata Siniscalco
University of Turin

01 — Overview

POLLEN13K is the first large-scale pollen grain image dataset, comprising more than 13,000 labeled objects extracted from aerobiological microscope images. It was built to support the development and evaluation of automatic pollen grain classification systems using modern machine learning and deep learning techniques.

Aerobiology — the study of airborne biological particles — plays a critical role in medicine, biology, and agronomy. Reliable estimation of airborne pollen abundance informs allergy risk assessment, plant phenology monitoring, and crop yield forecasting. However, manual analysis by qualified experts is time-consuming and limits scalability. POLLEN13K addresses this bottleneck by providing a curated, expert-validated resource for training and benchmarking automated classifiers.

13,353 Labeled Objects
5 Classes
84×84 Resolution (px)
RGB Color Images

02 — Dataset Classes

Under the guidance of aerobiology experts, all segmented objects were manually assigned to one of five categories. The Cupressaceae class contains very few examples (43 objects) and was excluded from the baseline classification experiments due to data imbalance.

Label Class Objects Notes
1 Corylus avellana (well-developed) 1,850 Normal hazel pollen grains
2 Corylus avellana (anomalous) 903 Morphologically aberrant hazel grains
3 Alnus (well-developed) 9,558 Alder pollen — dominant class
4 Debris 999 Air bubbles, dust, fungal spores, non-pollen objects
5 Cupressaceae 43 Cypress family pollen (low sample count)

Each object is provided as a color RGB image (84×84 px), together with its corresponding binary segmentation mask and a version with a uniform green background — facilitating experiments with different input modalities.

03 — Comparison with Existing Datasets

POLLEN13K substantially surpasses previous publicly available pollen datasets in terms of the number of annotated objects, making it uniquely suited for training data-hungry deep learning models.

Dataset Objects Image Type Resolution
Duller's Pollen Dataset 630 Grayscale 25×25
POLEN23E 805 Color ≥250 px/dim
Ranzato et al. 3,686 Color 1024×1024 (multi-grain)
POLLEN13K (Ours) >13,000 Color 84×84

04 — Data Acquisition & Segmentation

Airborne samples were collected on adhesive tapes mounted on a rotating drum (moving at 2 mm h⁻¹ under a suction hole). Pollen grains were inspected on daily segments using a Leitz Diaplan bright-field microscope with a 5 MP CMOS sensor. Pollen walls were stained with a mounting medium containing basic fuchsin.

Acquired images contained heavy background noise from debris, dust, fungal spores, and air bubbles introduced during mounting. A dedicated segmentation pipeline was developed to reliably locate and extract individual objects.

1
Mean Shift Filtering Flattens color texture while preserving object boundaries for downstream thresholding.
2
Background Smoothing & Otsu Thresholding Grayscale conversion, binary thresholding (threshold = 127), and removal of connected components smaller than 500 px.
3
Gaussian Blur & Contour Highlighting An 11×11 Gaussian kernel blurs background detail; detected contours are colored yellow to separate foreground from background.
4
RGB → HSV Conversion & Morphological Operations Closing and dilation (3×3 kernel) reduce noise; flood fill distinguishes foreground objects from background.
5
Adaptive Thresholding & Mask Refinement Gaussian adaptive threshold (block size 77, C = 0) followed by connected-component filtering (>150 px) and dilation (5 iterations) to produce clean binary masks.
6
Expert Validation & Labelling Aerobiology experts reviewed and labelled all segmented objects; 63 ambiguous instances (overlapping pollen and bubbles) were discarded.

05 — Baseline Classification Results

To establish baseline performance, several standard machine learning classifiers were evaluated using HOG and LBP feature descriptors, followed by experiments with deep convolutional neural networks. The dataset was split 85% / 15% for training and testing, using stratified sampling to preserve class proportions. The weighted F1 score was chosen as the primary metric to account for class imbalance.

Among classical approaches, RBF SVM with HOG features achieved the best result (accuracy: 86.6%, F1: 0.857). Among deep learning models, AlexNet trained on an augmented dataset (combining green-background and noisy-background images) reached an average F1 score of 0.87. SmallerVGGNet achieved 0.85 on the same augmented set. The most common misclassification occurred between well-developed Corylus avellana and Alnus grains, which share similar surface textures.

06 — Citation

If you use POLLEN13K in your research, we kindly ask that you cite the relevant publications listed below. Proper attribution helps sustain open dataset initiatives and supports the research community.
Primary Dataset Paper
@inproceedings{battiato2020pollen13k, title = {Pollen13k: A large scale microscope pollen grain image dataset}, author = {Battiato, Sebastiano and Ortis, Alessandro and Trenta, Francesca and Ascari, Lorenzo and Politi, Mara and Siniscalco, Consolata}, booktitle = {2020 IEEE International Conference on Image Processing (ICIP)}, pages = {2456--2460}, year = {2020}, organization = {IEEE} }
Detection & Classification (CVPR Workshop 2020)
@inproceedings{battiato2020detection, title = {Detection and classification of pollen grain microscope images}, author = {Battiato, Sebastiano and Ortis, Alessandro and Trenta, Francesca and Ascari, Lorenzo and Politi, Mara and Siniscalco, Consolata}, booktitle = {Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops}, pages = {980--981}, year = {2020} }
Classification Challenge Report (ICPR 2021)
@inproceedings{battiato2021pollen, title = {Pollen Grain Classification Challenge 2020: Challenge Report}, author = {Battiato, Sebastiano and Guarnera, Francesco and Ortis, Alessandro and Trenta, Francesca and Ascari, Lorenzo and Siniscalco, Consolata and De Gregorio, Tommaso and Su{\'a}rez, Eloy}, booktitle = {International Conference on Pattern Recognition}, pages = {469--479}, year = {2021}, organization = {Springer} }
Fine-Grained Classification (CAIP 2021)
@inproceedings{trenta2021fine, title = {Fine-grained image classification for pollen grain microscope images}, author = {Trenta, Francesca and Ortis, Alessandro and Battiato, Sebastiano}, booktitle = {International Conference on Computer Analysis of Images and Patterns}, pages = {341--351}, year = {2021}, organization = {Springer} }

07 — Acknowledgements

This research was carried out in collaboration with Ferrero HCo, which financed the project and enabled the collection of aerobiological samples from hazelnut plantations. The authors thank the aerobiology experts who contributed to sample collection, labelling, and validation.