Pollen Grain Recognition Using Deep Learning

OLEAtool: An open-source software for morphopalynological research in Olea europaea L. pollen

In this paper we present OLEAtool, a new software tool for palynological research to facilitate morphological analysis and measurements of Olea pollen. OLEAtool is a macro extension for use with ImageJ, an open-access and freely available image analysis software, and was developed as a component of the OLEA-project. This larger project examines olive tree expansion and mosaic landscape formation on the Balearic Islands. Pollen analysis of both fossil and modern grains has been proven useful for characterizing cultivars and therefore an important method for studying olive tree cultivation in the Mediterranean. However, these methods still struggle with distinguishing between wild and cultivated varieties. Traditional morphological analysis of pollen grains can be a difficult and time-consuming task. However, OLEAtool dramatically increases the speed of collecting data on pollen grains, expands the number of variables an analyst can measure, and greatly enhances the replicability of morphological analysis.

Deductive automated pollen classification in environmental samples via exploratory deep learning and imaging flow cytometry

Pollen and tracheophyte spores are ubiquitous environmental indicators at local and global scales. Palynology is typically performed manually by microscopic analysis; a specialised and time-consuming task limited in taxonomical precision and sampling frequency, therefore restricting data quality used to inform climate change and pollen forecasting models. We build on the growing work using AI (artificial intelligence) for automated pollen classification to design a flexible network that can deal with the uncertainty of broad-scale environmental applications. We combined imaging flow cytometry with Guided Deep Learning to identify and accurately categorise pollen in environmental samples; here, pollen grains captured within c. 5500 Cal yr BP old lake sediments. Our network discriminates not only pollen included in training libraries to the species level but, depending on the sample, can classify previously unseen pollen to the likely phylogenetic order, family and even genus. Our approach offers valuable insights into the development of a widely transferable, rapid and accurate exploratory tool for pollen classification in 'real-world' environmental samples with improved accuracy over pure deep learning techniques. This work has the potential to revolutionise many aspects of palynology, allowing a more detailed spatial and temporal understanding of pollen in the environment with improved taxonomical resolution.

Weakly Supervised Collaborative Learning for Airborne Pollen Segmentation and Classification from SEM Images

Existing pollen identification methods heavily rely on the scale and quality of pollen images. However, there are many impurities in real-world SEM images that should be considered. This paper proposes a collaborative learning method to jointly improve the performance of pollen segmentation and classification in a weakly supervised manner. It first locates pollen regions from the raw images based on the detection model. To improve the classification performance, we segmented the pollen grains through a pre-trained U-Net using unsupervised pollen contour features. The segmented pollen regions were fed into a deep convolutional neural network to obtain the activation maps, which were used to further refine the segmentation masks. In this way, both segmentation and classification models can be collaboratively trained, supervised by just pollen contour features and class-specific information. Extensive experiments on real-world datasets were conducted, and the results prove that our method effectively avoids impurity interference and improves pollen identification accuracy (86.6%) under the limited supervision (around 1000 images with image-level labels).

Simulation Palynologists for Pollinosis Prevention: A Progressive Learning of Pollen Localization and Classification for Whole Slide Images

Existing API approaches usually independently leverage detection or classification models to distinguish allergic pollens from Whole Slide Images (WSIs). However, palynologists tend to identify pollen grains in a progressive learning manner instead of the above one-stage straightforward way. They generally focus on two pivotal problems during pollen identification. (1) Localization: where are the pollen grains located? (2) Classification: which categories do these pollen grains belong to? To perfectly mimic the manual observation process of the palynologists, we propose a progressive method integrating pollen localization and classification to achieve allergic pollen identification from WSIs. Specifically, data preprocessing is first used to cut WSIs into specific patches and filter out blank background patches. Subsequently, we present the multi-scale detection model to locate coarse-grained pollen regions (targeting at "pollen localization problem") and the multi-classifiers combination to determine the fine-grained category of allergic pollens (targeting at "pollen classification problem"). Extensive experimental results have demonstrated the feasibility and effectiveness of our proposed method.

Electro-optical classification of pollen grains via microfluidics and machine learning

In aerobiological monitoring and agriculture there is a pressing need for accurate, label-free and automated analysis of pollen grains, in order to reduce the cost, workload and possible errors associated to traditional approaches. Methods: We propose a new multimodal approach that combines electrical sensing and optical imaging to classify pollen grains flowing in a microfluidic chip at a throughput of 150 grains per second. Electrical signals and synchronized optical images are processed by two independent machine learning-based classifiers, whose predictions are then combined to provide the final classification outcome. Results: The applicability of the method is demonstrated in a proof-of-concept classification experiment involving eight pollen classes from different taxa. The average balanced accuracy is 78.7 % for the electrical classifier, 76.7 % for the optical classifier and 84.2 % for the multimodal classifier. The accuracy is 82.8 % for the electrical classifier, 84.1 % for the optical classifier and 88.3 % for the multimodal classifier. Conclusion: The multimodal approach provides better classification results with respect to the analysis based on electrical or optical features alone. Significance: The proposed methodology paves the way for automated multimodal palynology. Moreover, it can be extended to other fields, such as diagnostics and cell therapy, where it could be used for label-free identification of cell populations in heterogeneous samples.

Neural networks for increased accuracy of allergenic pollen monitoring

Monitoring of airborne pollen concentrations provides an important source of information for the globally increasing number of hay fever patients. Airborne pollen is traditionally counted under the microscope, but with the latest developments in image recognition methods, automating this process has become feasible. A challenge that persists, however, is that many pollen grains cannot be distinguished beyond the genus or family level using a microscope. Here, we assess the use of Convolutional Neural Networks (CNNs) to increase taxonomic accuracy for airborne pollen. As a case study we use the nettle family (Urticaceae), which contains two main genera (Urtica and Parietaria) common in European landscapes which pollen cannot be separated by trained specialists. While pollen from Urtica species has very low allergenic relevance, pollen from several species of Parietaria is severely allergenic. We collect pollen from both fresh as well as from herbarium specimens and use these without the often used acetolysis step to train the CNN model. The models show that unacetolyzed Urticaceae pollen grains can be distinguished with > 98% accuracy. We then apply our model on before unseen Urticaceae pollen collected from aerobiological samples and show that the genera can be confidently distinguished, despite the more challenging input images that are often overlain by debris. Our method can also be applied to other pollen families in the future and will thus help to make allergenic pollen monitoring more specific.

Pollen wall patterns as a model for biological self-assembly

We are still far from being able to predict organisms' shapes purely from their genetic codes. While it is imperative to identify which encoded macromolecules contribute to a phenotype, determining how macromolecules self-assemble independently of the genetic code may be equally crucial for understanding shape development. Pollen grains are typically single-celled microgametophytes that have decorated walls of various shapes and patterns. The accumulation of morphological data and a comprehensive understanding of the wall development makes this system ripe for mathematical and physical modeling. Therefore, pollen walls are an excellent system for identifying both the genetic products and the physical processes that result in a huge diversity of extracellular morphologies. In this piece, I highlight the current understanding of pollen wall biology relevant for quantification studies and enumerate the modellable aspects of pollen wall patterning and specific approaches that one may take to elucidate how pollen grains build their beautifully patterned walls.