An official website of the United States government
Particle shape plays a critical role in governing the properties and behavior of granular materials. Despite advances in capturing and analyzing 3D particle shapes, these remain more demanding than 2D shape analysis due to the high computational costs and time-consuming nature of 3D imaging processes. Consequently, there is a growing interest in exploring potential correlations between 3D and 2D shapes, as this approach could potentially enable a reasonable estimation of a 3D shape from a 2D particle image, or at most, a couple of images. In response to this research interest, this study provides a thorough review of previous studies that have attempted to establish a correlation between 3D and 2D shape measures. A key finding from the extensive review is the high correlation between 2D perimeter circularity (cp) and Wadell’s true sphericity (S) defined in 3D, suggesting that a 3D shape can be estimated from the cp value in terms of S. To further substantiate the correlation between cp and S, this study analyzes approximately 400 mineral particle geometries available from an open-access data repository in both 3D and 2D. The analysis reveals a strong linear relationship between S and cp compared with other 2D shape descriptors broadly used in the research community. Furthermore, the limited variance in cp values indicates that cp is insensitive to changes in viewpoint, which indicates that fewer 2D images are needed. This finding offers a promising avenue for cost-effective and reliable 3D shape estimation using 2D particle images.
more »
« less
Phytoplankton carbon and nitrogen biomass estimates are robust to volume measurement method and growth environment
Abstract Phytoplankton biomass is routinely estimated using relationships between cell volume and carbon (C) and nitrogen (N) content that have been defined using diverse plankton that span orders of magnitude in size. Notably, volume has traditionally been estimated with geometric approximations of cell shape using cell dimensions from planar two-dimensional (2D) images, which requires assumptions about the third, depth dimension. Given advances in image processing, we examined how cell volumes determined from three-dimensional (3D), confocal images affected established relationships between phytoplankton cell volume and C and N content. Additionally, we determined that growth conditions could result in 30–40% variation in cellular N and C. 3D phytoplankton cell volume measurements were on average 15% greater than the geometric approximations from 2D images. Volume method variation was minimal compared to both intraspecific variation in volumes (~30%) and the 50-fold variation in elemental density among species. Consequently, C:vol and N:vol relationships were unaltered by volume measurement method and growth environment. Recent advances in instrumentation, including those for at sea and autonomous applications can be used to estimate plankton biomass directly. Going forward, we recommend instrumentation that permits species identification alongside size and shape characteristics for plankton biomass estimates.
more »
« less
- PAR ID:
- 10226078
- Date Published:
- Journal Name:
- Journal of Plankton Research
- Volume:
- 43
- Issue:
- 2
- ISSN:
- 0142-7873
- Page Range / eLocation ID:
- 103 to 112
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
Abstract. Oligotrophic regions play a central role in global biogeochemical cycles, with microbial communities in these areas representing an important term in global carbon budgets. While the general structure of microbial communities has been well documented in the global ocean, some remote regions such as the western tropical South Pacific (WTSP) remain fundamentally unexplored. Moreover, the biotic and abiotic factors constraining microbial abundances and distribution remain not well resolved. In this study, we quantified the spatial (vertical and horizontal) distribution of major microbial plankton groups along a transect through the WTSP during the austral summer of 2015, capturing important autotrophic and heterotrophic assemblages including cytometrically determined abundances of non-pigmented protists (also called flagellates). Using environmental parameters (e.g., nutrients and light availability) as well as statistical analyses, we estimated the role of bottom–up and top–down controls in constraining the structure of the WTSP microbial communities in biogeochemically distinct regions. At the most general level, we found a typical tropical structure, characterized by a shallow mixed layer, a clear deep chlorophyll maximum at all sampling sites, and a deep nitracline. Prochlorococcus was especially abundant along the transect, accounting for 68±10.6% of depth-integrated phytoplankton biomass. Despite their relatively low abundances, picophytoeukaryotes (PPE) accounted for up to 26±11.6% of depth-integrated phytoplankton biomass, while Synechococcus accounted for only 6±6.9%. Our results show that the microbial community structure of the WTSP is typical of highly stratified regions, and underline the significant contribution to total biomass by PPE populations. Strong relationships between N2 fixation rates and plankton abundances demonstrate the central role of N2 fixation in regulating ecosystem processes in the WTSP, while comparative analyses of abundance data suggest microbial community structure to be increasingly regulated by bottom–up processes under nutrient limitation, possibly in response to shifts in abundances of high nucleic acid bacteria (HNA).more » « less
-
Collecting large-scale medical datasets with fully annotated samples for training of deep networks is prohibitively expensive, especially for 3D volume data. Recent breakthroughs in self-supervised learning (SSL) offer the ability to overcome the lack of labeled training samples by learning feature representations from unlabeled data. However, most current SSL techniques in the medical field have been designed for either 2D images or 3D volumes. In practice, this restricts the capability to fully leverage unlabeled data from numerous sources, which may include both 2D and 3D data. Additionally, the use of these pre-trained networks is constrained to downstream tasks with compatible data dimensions.In this paper, we propose a novel framework for unsupervised joint learning on 2D and 3D data modalities. Given a set of 2D images or 2D slices extracted from 3D volumes, we construct an SSL task based on a 2D contrastive clustering problem for distinct classes. The 3D volumes are exploited by computing vectored embedding at each slice and then assembling a holistic feature through deformable self-attention mechanisms in Transformer, allowing incorporating long-range dependencies between slices inside 3D volumes. These holistic features are further utilized to define a novel 3D clustering agreement-based SSL task and masking embedding prediction inspired by pre-trained language models. Experiments on downstream tasks, such as 3D brain segmentation, lung nodule detection, 3D heart structures segmentation, and abnormal chest X-ray detection, demonstrate the effectiveness of our joint 2D and 3D SSL approach. We improve plain 2D Deep-ClusterV2 and SwAV by a significant margin and also surpass various modern 2D and 3D SSL approaches.more » « less
-
Abstract The Hawaii Ocean Time‐series (HOT) at Station ALOHA (22.75°N, 158°W) in the North Pacific Subtropical Gyre (NPSG) serves as a critical vantage point for observing plankton biomass production and its ecological implications. However, the HOT program's near‐monthly sampling frequency does not capture shorter time scale variability in phytoplankton populations. To address this gap, we deployed the SeaFlow flow cytometer for continuous monitoring during HOT cruises from 2014 to 2021. This approach allowed us to examine variations in the surface abundance and cell carbon content of specific phytoplankton groups: the cyanobacteriaProchlorococcus,Synechococcus, andCrocosphaeraas well as a range of small eukaryotic phytoplankton ( 5 μm). Our data showed that daily to monthly variability inProchlorococcusandSynechococcusabundance matches seasonal and interannual variability, while small eukaryotic phytoplankton andCrocosphaerashowed the highest seasonal and interannual fluctuations. The study also found that eukaryotic phytoplankton andCrocosphaerahad higher median cellular growth rates (0.076 and , respectively) compared toProchlorococcusandSynechococcus(0.037 and , respectively). These variances in abundance and growth rates indicate that shifts in the community structure significantly impact primary productivity in the NPSG. Our results underscore the importance of daily to monthly phytoplankton dynamics in ecosystem function and carbon cycling.more » « less
-
Three-dimensional (3D) discrete dislocation dynamics simulations are used to analyze the size effect on the fractal dimension of two-dimensional (2D) and 3D dislocation microstructure. 2D dislocation structures are analyzed first, and the calculated fractal dimension ( n 2 ) is found to be consistent with experimental results gleaned from transmission electron microscopy images. The value of n 2 is found to be close to unity for sizes smaller than 300 nm, and increases to a saturation value of ≈1.8 for sizes above approximately 10 microns. It is discovered that reducing the sample size leads to a decrease in the fractal dimension because of the decrease in the likelihood of forming strong tangles at small scales. Dislocation ensembles are found to exist in a more isolated way at the nano- and micro-scales. Fractal analysis is carried out on 3D dislocation structures and the 3D fractal dimension ( n 3 ) is determined. The analysis here shows that ( n 3 ) is significantly smaller than ( n 2 + 1 ) of 2D projected dislocations in all considered sizes.more » « less
- Free Publicly Accessible Full Text
-
Accepted Manuscript1.0
- Journal Article:
- https://doi.org/10.1093/plankt/fbab014
