Efficient Wheat Head Segmentation with Minimal Annotation: A Generative Approach.

Deep learning models have been used for a variety of image processing tasks. However, most of these models are developed through supervised learning approaches, which rely heavily on the availability of large-scale annotated datasets. Developing such datasets is tedious and expensive. In the absence of an annotated dataset, synthetic data can be used for model development; however, due to the substantial differences between simulated and real data, a phenomenon referred to as domain gap, the resulting models often underperform when applied to real data. In this research, we aim to address this challenge by first computationally simulating a large-scale annotated dataset and then using a generative adversarial network (GAN) to fill the gap between simulated and real images. This approach results in a synthetic dataset that can be effectively utilized to train a deep-learning model. Using this approach, we developed a realistic annotated synthetic dataset for wheat head segmentation. This dataset was then used to develop a deep-learning model for semantic segmentation. The resulting model achieved a Dice score of 83.4% on an internal dataset and Dice scores of 79.6% and 83.6% on two external datasets from the Global Wheat Head Detection datasets. While we proposed this approach in the context of wheat head segmentation, it can be generalized to other crop types or, more broadly, to images with dense, repeated patterns such as those found in cellular imagery.

Predicting phenotypes from novel genomic markers using deep learning.

Genomic selection (GS) models use single nucleotide polymorphism (SNP) markers to predict phenotypes. However, these predictive models face challenges due to the high dimensionality of genome-wide SNP marker data. Thanks to recent breakthroughs in DNA sequencing and decreased sequencing cost, the study of novel genomic variants such as structural variations (SVs) and transposable elements (TEs) become increasingly prevalent. In this article, we develop a deep convolutional neural network model, NovGMDeep, to predict phenotypes using SVs and TEs markers for GS. The proposed model is trained and tested on samples of Arabidopsis thaliana and Oryza sativa using k-fold cross-validation. The prediction accuracy is evaluated using Pearson's Correlation Coefficient (PCC), mean absolute error (MAE) and SD of MAE. The predicted results showed higher correlation when the model is trained with SVs and TEs than with SNPs. NovGMDeep also has higher prediction accuracy when comparing with conventional statistical models. This work sheds light on the unappreciated function of SVs and TEs in genotype-to-phenotype associations, as well as their extensive significance and value in crop development.

Integument colour change: Tracking delayed growth of Oppia nitens as a sub-lethal indicator of soil toxicity.

Growth is an important toxicity end-point in ecotoxicology but is rarely used in soil ecotoxicological studies. Here, we assessed the growth change of Oppia nitens when exposed to reference and heavy metal toxicants. To assess mite growth, we developed an image analysis methodology to measure colour spectrum changes of the mite integument at the final developmental stage, as a proxy for growth change. We linked the values of red, green, blue, key-black, and light colour of mites to different growth stages. Based on this concept, we assessed the growth change of mites exposed to cadmium, copper, zinc, lead, boric acid, or phenanthrene at sublethal concentrations in LUFA 2.2 soil for 14 days. Sublethal effects were detected after 7 days of exposure. The growth of O. nitens was more sensitive than survival and reproduction when exposed to copper (EC50growth = 1360 mg/kg compared to EC50reproduction = 2896 mg/kg). Mite growth sensitivity was within the same order of magnitude to mite reproduction when exposed to zinc (EC50growth = 1785; EC50reproduction = 1562 mg/kg). At least 25% of sublethal effects of boric acid and phenanthrene were detected in the mites but growth was not impacted when O. nitens were exposed to lead. Consistent with previous studies, cadmium was the most toxic metal to O. nitens. The mite growth pattern was comparable to mite survival and reproduction from previous studies. Mite growth is a sensitive toxicity endpoint, ecologically relevant, fast, easy to detect, and can be assessed in a non-invasive fashion, thereby complimenting existing O. nitens testing protocols.

Semi-Self-Supervised Learning for Semantic Segmentation in Images with Dense Patterns.

Deep learning has shown potential in domains with large-scale annotated datasets. However, manual annotation is expensive, time-consuming, and tedious. Pixel-level annotations are particularly costly for semantic segmentation in images with dense irregular patterns of object instances, such as in plant images. In this work, we propose a method for developing high-performing deep learning models for semantic segmentation of such images utilizing little manual annotation. As a use case, we focus on wheat head segmentation. We synthesize a computationally annotated dataset-using a few annotated images, a short unannotated video clip of a wheat field, and several video clips with no wheat-to train a customized U-Net model. Considering the distribution shift between the synthesized and real images, we apply three domain adaptation steps to gradually bridge the domain gap. Only using two annotated images, we achieved a Dice score of 0.89 on the internal test set. When further evaluated on a diverse external dataset collected from 18 different domains across five countries, this model achieved a Dice score of 0.73. To expose the model to images from different growth stages and environmental conditions, we incorporated two annotated images from each of the 18 domains to further fine-tune the model. This increased the Dice score to 0.91. The result highlights the utility of the proposed approach in the absence of large-annotated datasets. Although our use case is wheat head segmentation, the proposed approach can be extended to other segmentation tasks with similar characteristics of irregularly repeating patterns of object instances.

Overview of Machine Learning Part 1: Fundamentals and Classic Approaches.

The extensive body of research and advances in machine learning (ML) and the availability of a large volume of patient data make ML a powerful tool for producing models with the potential for widespread deployment in clinical settings. This article provides an overview of the classic supervised and unsupervised ML methods as well as fundamental concepts required for understanding how to develop generalizable and high-performing ML applications. It also describes the important steps for developing a ML model and how decisions made in these steps affect model performance and ability to generalize.