Publication

Article Metrics

Citations


Online attention

Temporal refinement of 3D CNN semantic segmentations on 4D time-series of undersampled tomograms using hidden Markov models

DOI: 10.1038/s41598-021-02466-x DOI Help

Authors: Dimitrios Bellos (University of Nottingham; Diamond Light Source) , Mark Basham (Diamond Light Source; Rosalind Franklin Institute) , Tony Pridmore (University of Nottingham) , Andrew P. French (University of Nottingham)
Co-authored by industrial partner: No

Type: Journal Paper
Journal: Scientific Reports , VOL 11

State: Published (Approved)
Published: December 2021
Diamond Proposal Number(s): 9396

Open Access Open Access

Abstract: Recently, several convolutional neural networks have been proposed not only for 2D images, but also for 3D and 4D volume segmentation. Nevertheless, due to the large data size of the latter, acquiring a sufficient amount of training annotations is much more strenuous than in 2D images. For 4D time-series tomograms, this is usually handled by segmenting the constituent tomograms independently through time with 3D convolutional neural networks. Inter-volume information is therefore not utilized, potentially leading to temporal incoherence. In this paper, we attempt to resolve this by proposing two hidden Markov model variants that refine 4D segmentation labels made by 3D convolutional neural networks working on each time point. Our models utilize not only inter-volume information, but also the prediction confidence generated by the 3D segmentation convolutional neural networks themselves. To the best of our knowledge, this is the first attempt to refine 4D segmentations made by 3D convolutional neural networks using hidden Markov models. During experiments we test our models, qualitatively, quantitatively and behaviourally, using prespecified segmentations. We demonstrate in the domain of time series tomograms which are typically undersampled to allow more frequent capture; a particularly challenging problem. Finally, our dataset and code is publicly available.

Journal Keywords: Computer science; Scientific data; Software

Subject Areas: Information and Communication Technology


Instruments: I13-2-Diamond Manchester Imaging

Added On: 06/12/2021 11:46

Documents:
s41598-021-02466-x.pdf

Discipline Tags:

Information & Communication Technologies Artificial Intelligence Data processing

Technical Tags:

Imaging Tomography