Machine Learning for Analysis of Microscopy Images: A Practical Guide

Zinchuk, Vadim S.; Grossenbacher-Zinchuk, Olga

doi:10.1002/cpcb.101

Cited by 23 publications

(19 citation statements)

References 45 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Training data was obtained from four image stacks, selected from four different captures and two imaging sessions. Compared to the deep learning models sometimes used for semantic segmentation [ 41 ], the Trainable Weka approach requires only a very small selection of training data, although high quality results may require an iterative process involving assessment of representative segmentations, followed by updating the model with additional training data. Our macrophage model required two revision steps, primarily to add robustness against heterogeneity in noise and contrast between image captures.…”

Section: Resultsmentioning

confidence: 99%

“…The role of the image features, using algorithms provided by the ImageJ platform and the ImageScience plugin [ 23 ], resembles that of the earlier convolutional layers in the deep learning models such as U-Net [ 29 ] sometimes used for semantic segmentation [ 41 ]. However, using predefined image features radically reduces the cost of training in computational time and in the requirement for manually segmented training data, although the modeller must ensure that the selected features and scales capture sufficient spatial context for pixel classification.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

LLAMA: a robust and scalable machine learning pipeline for analysis of large scale 4D microscopy data: analysis of cell ruffles and filopodia

et al. 2021

View full text Add to dashboard Cite

Background With recent advances in microscopy, recordings of cell behaviour can result in terabyte-size datasets. The lattice light sheet microscope (LLSM) images cells at high speed and high 3D resolution, accumulating data at 100 frames/second over hours, presenting a major challenge for interrogating these datasets. The surfaces of vertebrate cells can rapidly deform to create projections that interact with the microenvironment. Such surface projections include spike-like filopodia and wave-like ruffles on the surface of macrophages as they engage in immune surveillance. LLSM imaging has provided new insights into the complex surface behaviours of immune cells, including revealing new types of ruffles. However, full use of these data requires systematic and quantitative analysis of thousands of projections over hundreds of time steps, and an effective system for analysis of individual structures at this scale requires efficient and robust methods with minimal user intervention. Results We present LLAMA, a platform to enable systematic analysis of terabyte-scale 4D microscopy datasets. We use a machine learning method for semantic segmentation, followed by a robust and configurable object separation and tracking algorithm, generating detailed object level statistics. Our system is designed to run on high-performance computing to achieve high throughput, with outputs suitable for visualisation and statistical analysis. Advanced visualisation is a key element of LLAMA: we provide a specialised tool which supports interactive quality control, optimisation, and output visualisation processes to complement the processing pipeline. LLAMA is demonstrated in an analysis of macrophage surface projections, in which it is used to i) discriminate ruffles induced by lipopolysaccharide (LPS) and macrophage colony stimulating factor (CSF-1) and ii) determine the autonomy of ruffle morphologies. Conclusions LLAMA provides an effective open source tool for running a cell microscopy analysis pipeline based on semantic segmentation, object analysis and tracking. Detailed numerical and visual outputs enable effective statistical analysis, identifying distinct patterns of increased activity under the two interventions considered in our example analysis. Our system provides the capacity to screen large datasets for specific structural configurations. LLAMA identified distinct features of LPS and CSF-1 induced ruffles and it identified a continuity of behaviour between tent pole ruffling, wave-like ruffling and filopodia deployment.

show abstract

Section: Resultsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

LLAMA: a robust and scalable machine learning pipeline for analysis of large scale 4D microscopy data: analysis of cell ruffles and filopodia

et al. 2021

View full text Add to dashboard Cite

show abstract

“…The role of the image features, using algorithms provided by the ImageJ platform and the ImageScience plugin (Meijering, 2015), resembles that of the earlier convolutional layers in the deep learning models sometimes used for semantic segmentation (Zinchuk & Grossenbacher-Zinchuk, 2020). However, using predefined image features radically reduces the cost of training in computational time and in the requirement for manually segmented training data, although the modeller must ensure that the selected features and scales capture sufficient spatial context for pixel classification.…”

Section: Methodsmentioning

confidence: 99%

“…Training data was obtained from four image stacks, selected from four different captures and two imaging sessions. Compared to the deep learning models sometimes used for semantic segmentation (Zinchuk & Grossenbacher-Zinchuk, 2020), the Trainable Weka approach requires only a very small selection of training data, although high quality results may require an iterative process involving assessment of representative segmentations, followed by updating the model with additional training data. A total of 16330 voxels were labelled with one of the 4 segmentation classes, equivalent to less than 0.1% of a single image stack cropped to the size of a macrophage (approx.…”

Section: Semantic Segmentationmentioning

confidence: 99%

LLAMA: a robust and scalable machine learning pipeline for analysis of cell surface projections in large scale 4D microscopy data

Koh

et al. 2020

Preprint

View full text Add to dashboard Cite

We present LLAMA, a pipeline for systematic analysis of terabyte scale 4D microscopy datasets. Analysis of individual biological structures in imaging at this scale requires efficient and robust methods which do not require human micromanagement or editing of outputs. To meet this challenge, we use a machine learning method for semantic segmentation, followed by a robust and configurable object separation and tracking algorithm, and the generation of detailed object level statistics. Advanced visualisation is a key element of LLAMA: we provide a specialised software tool which supports quality control and optimisation as well as visualisation of outputs. LLAMA was used in a quantitative analysis of macrophage surface membrane projections (filopodia, ruffles, tent-pole ruffles) examining the differential effects of two interventions: lipopolysaccharide (LPS) and macrophage colony stimulating factor (CSF-1). Distinct patterns of increased activity were identified. In addition, a continuity of behaviour was found between tent pole ruffling and wave-like ruffling, further defining the role of filopodia in ruffling.

show abstract

“…A review on future trends in microscopy around that time already commented that for complex visual tasks “a good deal of faith is now placed in electronic neural networks” [181] . Indeed, the use of ANNs caught on during the 1990s [182] , [183] , [184] and 2000s [185] , [186] , [187] , but as in biomedical imaging at large, deep learning began to be massively adopted for bioimage analysis only in recent years [188] , [189] , [190] , [191] , [192] , [193] , [194] . We briefly discuss some of the common tasks in bioimage analysis ( Fig.…”

Section: Deep Learning For Bioimage Analysismentioning

confidence: 99%

A bird’s-eye view of deep learning in bioimage analysis

Meijering

2020

Computational and Structural Biotechnology Journal

112

View full text Add to dashboard Cite

Deep learning of artificial neural networks has become the de facto standard approach to solving data analysis problems in virtually all fields of science and engineering. Also in biology and medicine, deep learning technologies are fundamentally transforming how we acquire, process, analyze, and interpret data, with potentially far-reaching consequences for healthcare. In this mini-review, we take a bird’s-eye view at the past, present, and future developments of deep learning, starting from science at large, to biomedical imaging, and bioimage analysis in particular.

show abstract

Machine Learning for Analysis of Microscopy Images: A Practical Guide

Cited by 23 publications

References 45 publications

LLAMA: a robust and scalable machine learning pipeline for analysis of large scale 4D microscopy data: analysis of cell ruffles and filopodia

LLAMA: a robust and scalable machine learning pipeline for analysis of large scale 4D microscopy data: analysis of cell ruffles and filopodia

LLAMA: a robust and scalable machine learning pipeline for analysis of cell surface projections in large scale 4D microscopy data

A bird’s-eye view of deep learning in bioimage analysis

Contact Info

Product

Resources

About