A Parallel Distributed-Memory Particle Method Enables Acquisition-Rate Segmentation of Large Fluorescence Microscopy Images

Afshar, Yaser; Sbalzarini, Ivo F.

doi:10.1371/journal.pone.0152528

Cited by 13 publications

(19 citation statements)

References 40 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…The source code of OpenFPM, virtual machines for various operating systems with a complete OpenFPM environment pre-installed, virtualized Docker containers, documentation, example applications, and tutorial videos are freely available from http://openfpm.mpi-cbg.de. We hope that the flexibility, free availability, performance, quality of documentation, and longterm support of OpenFPM will make it a standard platform for particles-only and hybrid particle-28 mesh simulations of discrete and continuous models on parallel computer hardware, as well as for non-simulation applications, such as evolutionary optimization strategies and particle-based image-analysis methods [83,73].…”

Section: Discussionmentioning

confidence: 99%

“…One of the main advantages of OpenFPM over other simulation frameworks is that OpenFPM can transparently handle spaces of arbitrary dimension. This enables simulations in higherdimensional spaces, such as the four-dimensional spaces used in lattice quantum chromodynamics [71,72], and it also enables parallelization of non-simulation applications that require 25 high-dimensional spaces, including image analysis algorithms [73] and Monte-Carlo sampling strategies [74]. A particular Monte-Carlo sampler used for stochastic real-valued optimization is the Covariance-Matrix-Adaptation Evolution Strategy (CMA-ES) [75,76].…”

Section: Particle-swarm Covariance-matrix-adaptation Evolution Stratementioning

confidence: 99%

See 1 more Smart Citation

OpenFPM: A scalable open framework for particle and particle-mesh codes on parallel computers

Incardona

Leo

Zaluzhnyi

et al. 2019

Computer Physics Communications

Self Cite

View full text Add to dashboard Cite

Scalable and efficient numerical simulations continue to gain importance, as computation is firmly established as the third pillar of discovery, alongside theory and experiment. Meanwhile, the performance of computing hardware grows through increasing heterogeneous parallelism, enabling simulations of ever more complex models. However, efficiently implementing scalable codes on heterogeneous, distributed hardware systems becomes the bottleneck. This bottleneck can be alleviated by intermediate software layers that provide higher-level abstractions closer to the problem domain, hence allowing the computational scientist to focus on the simulation. Here, we present OpenFPM, an open and scalable framework that provides an abstraction layer for numerical simulations using particles and/or meshes. OpenFPM provides transparent and scalable infrastructure for shared-memory and distributed-memory implementations of particlesonly and hybrid particle-mesh simulations of both discrete and continuous models, as well as non-simulation codes. This infrastructure is complemented with portable implementations of frequently used numerical routines, as well as interfaces to third-party libraries. We present the architecture and design of OpenFPM, detail the underlying abstractions, and benchmark the framework in applications ranging from Smoothed-Particle Hydrodynamics (SPH) to Molecular Dynamics (MD), Discrete Element Methods (DEM), Vortex Methods, stencil codes, highdimensional Monte Carlo sampling (CMA-ES), and Reaction-Diffusion solvers, comparing it to the current state of the art and existing software frameworks.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Particle-swarm Covariance-matrix-adaptation Evolution Stratementioning

confidence: 99%

OpenFPM: A scalable open framework for particle and particle-mesh codes on parallel computers

Incardona

Leo

Zaluzhnyi

et al. 2019

Computer Physics Communications

Self Cite

View full text Add to dashboard Cite

show abstract

“…Their method was tested with human glioblastoma cell culture IF images and obtained high agreement with their manual GT. Finally, in a very recent article, Afshar et al [8] proposed a distributed-memory algorithm for segmentation where each boundary pixel is modeled as a particle system that evolves in time. Authors showed good results in synthetic images and in cell culture IF images.…”

Section: Discussionmentioning

confidence: 99%

“…Several authors developed algorithms to automate IF analysis using unsupervised methods, mostly based on deformable contours and region-based approaches [2, 3, 4, 5, 6, 7, 8] and fewer focused on signal processing [9, 10]. Despite a great deal of publications in cell analyses, most of the methods for detecting and segmenting cells in IF images have been tailored to work with images derived from cell cultures or experimental animals, in which several parameters can be controlled to render IF images with a rather smooth, clean background and high contrast.…”

Section: Introductionmentioning

confidence: 99%

Automating cell detection and classification in human brain fluorescent microscopy images using dictionary learning and sparse coding

Alegro

Theofilas

Nguy

et al. 2017

Journal of Neuroscience Methods

View full text Add to dashboard Cite

Background Immunofluorescence (IF) plays a major role in quantifying protein expression in situ and understanding cell function. It is widely applied in assessing disease mechanisms and in drug discovery research. Automation of IF analysis can transform studies using experimental cell models. However, IF analysis of postmortem human tissue relies mostly on manual interaction, often subjected to low-throughput and prone to error, leading to low inter and intra-observer reproducibility. Human postmortem brain samples challenges neuroscientists because of the high level of autofluorescence caused by accumulation of lipofuscin pigment during aging, hindering systematic analyses. We propose a method for automating cell counting and classification in IF microscopy of human postmortem brains. Our algorithm speeds up the quantification task while improving reproducibility. New method Dictionary learning and sparse coding allow for constructing improved cell representations using IF images. These models are input for detection and segmentation methods. Classification occurs by means of color distances between cells and a learned set. Results Our method successfully detected and classified cells in 49 human brain images. We evaluated our results regarding true positive, false positive, false negative, precision, recall, false positive rate and F1 score metrics. We also measured user-experience and time saved compared to manual countings. Comparison with existing methods We compared our results to four open-access IF-based cell-counting tools available in the literature. Our method showed improved accuracy for all data samples. Conclusion The proposed method satisfactorily detects and classifies cells from human postmortem brain IF images, with potential to be generalized for applications in other counting tasks.

show abstract

“…The pipeline shows efficient parallel scaling (Amdahl's Law, parallel fraction = 0.95) on up to 47 cores, achieving data rates of up to 1400 MB/second (SFigure 35). This enables real-time conversion of images to the APR, as it is faster than the acquisition rate of microscopes (32,33).…”

Section: Benchmarks On Real Datamentioning

confidence: 99%

Forget Pixels: Adaptive Particle Representation of Fluorescence Microscopy Images

Cheeseman

Günther

Susik

et al. 2018

Preprint

Self Cite

View full text Add to dashboard Cite

Modern microscopy modalities create a data deluge with gigabytes of data generated each second, or terabytes per day. Storing and processing these data is a severe bottleneck. We argue that this is an artifact of the images being represented on pixels. To address the root of the problem, we here propose the Adaptive Particle Representation (APR) as an image-content-aware representation of fluorescence microscopy images. The APR replaces pixel images to overcome computational and memory bottlenecks in storage and processing pipelines for studying spatiotemporal processes in biology using fluorescence microscopy. We present the ideas, concepts, and algorithms and validate them using noisy 3D image data. We show how the APR adapts to the information content of an image without reducing image quality. We then show that the 1 . CC-BY-NC-ND 4.0 International license peer-reviewed) is the author/funder. It is made available under aThe copyright holder for this preprint (which was not . http://dx.doi.org/10.1101/263061 doi: bioRxiv preprint first posted online Feb. 9, 2018; adaptivity of the APR provides orders of magnitude benefits across a range of image-processing tasks. Therefore, the APR provides a simple, extendable, and efficient content-aware representation of images that could be useful for many imaging modalities in order to relax current data and processing bottlenecks.(less than 15000 words, 6 Figures)Introduction New developments in fluorescence microscopy (1-3), labeling chemistry (4), and genetics (5) provide the potential to capture and track biological structures at high resolution in both space and time. Such data is vital for understanding many spatiotemporal processes in biology (6). Unfortunately, fluorescence microscopes do not directly output the shapes and locations of objects through time. Instead, they produce raw data, potentially terabytes of 3D images (7), from which the desired spatiotemporal information must be extracted by image processing. Handling the large image data and extracting information from the raw microscopy images presents the main bottleneck (7-9). We propose that at the core of the problem is not the amount of information contained in the images, but how the data encodes this information -usually as pixels on a uniform grid.The uniform grids of pixels in the images contain information on labeled objects quantifying local intensity variations of the fluorescence signal. These local intensity variations are a measurement of the spatial localization of fluorescent molecules. Inferring information about the shape and location of labeled structures is complicated by structures of equal interest showing different imaged intensities, hence defining locally different scales of intensity variation across the image. The wide range of spatial and temporal scales in biological processes often requires imaging a large field of view at both high spatial and temporal resolution. This uniformly high resolution exacerbates the data problem and amplifies the processing bottleneck.

show abstract

A Parallel Distributed-Memory Particle Method Enables Acquisition-Rate Segmentation of Large Fluorescence Microscopy Images

Cited by 13 publications

References 40 publications

OpenFPM: A scalable open framework for particle and particle-mesh codes on parallel computers

OpenFPM: A scalable open framework for particle and particle-mesh codes on parallel computers

Automating cell detection and classification in human brain fluorescent microscopy images using dictionary learning and sparse coding

Forget Pixels: Adaptive Particle Representation of Fluorescence Microscopy Images

Contact Info

Product

Resources

About