Milling Manuel, Rampp Simon D N, Triantafyllopoulos Andreas, Plaza Maria P, Brunner Jens O, Traidl-Hoffmann Claudia, Schuller Björn W, Damialis Athanasios
CHI - Chair of Health Informatics, MRI, Technical University of Munich, Munich, Germany.
MCML-Munich Center for Machine Learning, Germany.
Heliyon. 2025 Jan 3;11(2):e41656. doi: 10.1016/j.heliyon.2025.e41656. eCollection 2025 Jan 30.
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.
基于深度学习的花粉粒分类一直是实现空气中花粉自动监测的主要驱动力。然而，尽管有大量可用的数据集，但人们很少花精力去研究哪些方面对（通常类似于黑箱的）花粉分类方法构成最大挑战。为了阐明这个问题，我们基于显微镜图像上最大的自动生成的花粉粒数据集之一的可能性进行了样本级难度分析，并研究了某些空气中样本和特定花粉分类群给深度学习算法带来特殊问题的原因。在此得出结论，主要挑战在于：A）单个图像中多个花粉粒（部分）同时出现；B）显微镜图像的二维捕捉导致特定标记被遮挡；C）对于某些分类群，普遍缺乏显著、独特的特征。我们的代码可在https://github.com/millinma/SDPollen上公开获取。
