Automating airborne pollen classification: Identifying and interpreting hard samples for classifiers
Deep-learning-based classification of pollen grains has been a major driver towards automatic monitoring of airborne pollen. Yet, despite an abundance of available datasets, little effort has been spent to investigate which aspects pose the biggest challenges to the (often black-box- resembling) pol...
Saved in:
| Main Authors: | , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Elsevier
2025-01-01
|
| Series: | Heliyon |
| Subjects: | |
| Online Access: | http://www.sciencedirect.com/science/article/pii/S2405844025000362 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1832573143382753280 |
|---|---|
| author | Manuel Milling Simon D.N. Rampp Andreas Triantafyllopoulos Maria P. Plaza Jens O. Brunner Claudia Traidl-Hoffmann Björn W. Schuller Athanasios Damialis |
| author_facet | Manuel Milling Simon D.N. Rampp Andreas Triantafyllopoulos Maria P. Plaza Jens O. Brunner Claudia Traidl-Hoffmann Björn W. Schuller Athanasios Damialis |
| author_sort | Manuel Milling |
| collection | DOAJ |
| description | Deep-learning-based classification of pollen grains has been a major driver towards automatic monitoring of airborne pollen. Yet, despite an abundance of available datasets, little effort has been spent to investigate which aspects pose the biggest challenges to the (often black-box- resembling) pollen classification approaches. To shed some light on this issue, we conducted a sample-level difficulty analysis based on the likelihood for one of the largest automatically-generated datasets of pollen grains on microscopy images and investigated the reason for which certain airborne samples and specific pollen taxa pose particular problems to deep learning algorithms. It is here concluded that the main challenges lie in A) the (partly) co-occurring of multiple pollen grains in a single image, B) the occlusion of specific markers through the 2D capturing of microscopy images, and C) for some taxa, a general lack of salient, unique features. Our code is publicly available under https://github.com/millinma/SDPollen |
| format | Article |
| id | doaj-art-50c2b66854a8487aad85965ca24d9d91 |
| institution | Kabale University |
| issn | 2405-8440 |
| language | English |
| publishDate | 2025-01-01 |
| publisher | Elsevier |
| record_format | Article |
| series | Heliyon |
| spelling | doaj-art-50c2b66854a8487aad85965ca24d9d912025-02-02T05:27:56ZengElsevierHeliyon2405-84402025-01-01112e41656Automating airborne pollen classification: Identifying and interpreting hard samples for classifiersManuel Milling0Simon D.N. Rampp1Andreas Triantafyllopoulos2Maria P. Plaza3Jens O. Brunner4Claudia Traidl-Hoffmann5Björn W. Schuller6Athanasios Damialis7CHI – Chair of Health Informatics, MRI, Technical University of Munich, Munich, Germany; MCML–Munich Center for Machine Learning, Germany; EIHW – Chair of Embedded Intelligence for Health Care & Wellbeing, University of Augsburg, Augsburg, GermanyEIHW – Chair of Embedded Intelligence for Health Care & Wellbeing, University of Augsburg, Augsburg, GermanyCHI – Chair of Health Informatics, MRI, Technical University of Munich, Munich, Germany; MCML–Munich Center for Machine Learning, Germany; EIHW – Chair of Embedded Intelligence for Health Care & Wellbeing, University of Augsburg, Augsburg, GermanyInstitute of Environmental Medicine and Integrative Health, Faculty of Medicine, University Clinic of Augsburg & University of Augsburg, Augsburg, Germany; Institute of Environmental Medicine, Helmholtz Center Munich, German Research Center for Environmental Health, GermanyFaculty of Business and Economics, and Faculty of Medicine, University of Augsburg, Augsburg, Germany; Department of Technology, Management, and Economics, Technical University of Denmark, Denmark; Next Generation Technology, Region Zealand, DenmarkInstitute of Environmental Medicine and Integrative Health, Faculty of Medicine, University Clinic of Augsburg & University of Augsburg, Augsburg, Germany; Institute of Environmental Medicine, Helmholtz Center Munich, German Research Center for Environmental Health, Germany; Christine Kühne Center for Allergy Research and Education, Davos, SwitzerlandCHI – Chair of Health Informatics, MRI, Technical University of Munich, Munich, Germany; MCML–Munich Center for Machine Learning, Germany; EIHW – Chair of Embedded Intelligence for Health Care & Wellbeing, University of Augsburg, Augsburg, Germany; MDSI–Munich Data Science Institute, Germany; GLAM–the Group on Language, Audio, & Music, Imperial College London, London, UKInstitute of Environmental Medicine and Integrative Health, Faculty of Medicine, University Clinic of Augsburg & University of Augsburg, Augsburg, Germany; Terrestrial Ecology and Climate Change, Department of Ecology, School of Biology, Faculty of Sciences, Aristotle University of Thessaloniki, Thessaloniki, Greece; Corresponding author. Terrestrial Ecology and Climate Change, Department of Ecology, School of Biology, Faculty of Sciences, Aristotle University of Thessaloniki, Thessaloniki, Greece.Deep-learning-based classification of pollen grains has been a major driver towards automatic monitoring of airborne pollen. Yet, despite an abundance of available datasets, little effort has been spent to investigate which aspects pose the biggest challenges to the (often black-box- resembling) pollen classification approaches. To shed some light on this issue, we conducted a sample-level difficulty analysis based on the likelihood for one of the largest automatically-generated datasets of pollen grains on microscopy images and investigated the reason for which certain airborne samples and specific pollen taxa pose particular problems to deep learning algorithms. It is here concluded that the main challenges lie in A) the (partly) co-occurring of multiple pollen grains in a single image, B) the occlusion of specific markers through the 2D capturing of microscopy images, and C) for some taxa, a general lack of salient, unique features. Our code is publicly available under https://github.com/millinma/SDPollenhttp://www.sciencedirect.com/science/article/pii/S2405844025000362Pollen recognitionSample difficulty analysisDeep learning |
| spellingShingle | Manuel Milling Simon D.N. Rampp Andreas Triantafyllopoulos Maria P. Plaza Jens O. Brunner Claudia Traidl-Hoffmann Björn W. Schuller Athanasios Damialis Automating airborne pollen classification: Identifying and interpreting hard samples for classifiers Heliyon Pollen recognition Sample difficulty analysis Deep learning |
| title | Automating airborne pollen classification: Identifying and interpreting hard samples for classifiers |
| title_full | Automating airborne pollen classification: Identifying and interpreting hard samples for classifiers |
| title_fullStr | Automating airborne pollen classification: Identifying and interpreting hard samples for classifiers |
| title_full_unstemmed | Automating airborne pollen classification: Identifying and interpreting hard samples for classifiers |
| title_short | Automating airborne pollen classification: Identifying and interpreting hard samples for classifiers |
| title_sort | automating airborne pollen classification identifying and interpreting hard samples for classifiers |
| topic | Pollen recognition Sample difficulty analysis Deep learning |
| url | http://www.sciencedirect.com/science/article/pii/S2405844025000362 |
| work_keys_str_mv | AT manuelmilling automatingairbornepollenclassificationidentifyingandinterpretinghardsamplesforclassifiers AT simondnrampp automatingairbornepollenclassificationidentifyingandinterpretinghardsamplesforclassifiers AT andreastriantafyllopoulos automatingairbornepollenclassificationidentifyingandinterpretinghardsamplesforclassifiers AT mariapplaza automatingairbornepollenclassificationidentifyingandinterpretinghardsamplesforclassifiers AT jensobrunner automatingairbornepollenclassificationidentifyingandinterpretinghardsamplesforclassifiers AT claudiatraidlhoffmann automatingairbornepollenclassificationidentifyingandinterpretinghardsamplesforclassifiers AT bjornwschuller automatingairbornepollenclassificationidentifyingandinterpretinghardsamplesforclassifiers AT athanasiosdamialis automatingairbornepollenclassificationidentifyingandinterpretinghardsamplesforclassifiers |