Segmentation of cone-beam CT using a hidden Markov random field with informative priors

Moores, Matthew T.; Hargrave, Catriona; Harden, Fiona; Mengersen, Kerrie

doi:10.1088/1742-6596/489/1/012076

Cited by 9 publications

(9 citation statements)

References 8 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…While being widely used in practice (e.g. Hurn et al, ; Alfò et al, ; François et al, ; Moores et al, ), the model is representative of the computational difficulties of hidden Gibbs random field. The model sets a probability distribution on

X^{n} = {({}, 1, \dots, K)}^{n}

parametrized by a scalar ψ that adjusts the level of dependence between adjacent sites and whose Hamiltonian is given by

H (x | ψ, G) = - ψ \sum_{iG \sim j} 1 {x_{i} = x_{j}} .

The aforementioned sum

iG \sim j

ranges the set of edges of the graph

G

.…”

Section: Comparison Of Bic Approximationsmentioning

confidence: 99%

Hidden Gibbs random fields model selection using Block Likelihood Information Criterion

Stoehr

Marin

Pudlo

2016

Stat

View full text Add to dashboard Cite

Performing model selection between Gibbs random fields is a very challenging task. Indeed, due to the Markovian dependence structure, the normalizing constant of the fields cannot be computed using standard analytical or numerical methods. Furthermore, such unobserved fields cannot be integrated out and the likelihood evaluztion is a doubly intractable problem. This forms a central issue to pick the model that best fits an observed data. We introduce a new approximate version of the Bayesian Information Criterion. We partition the lattice into continuous rectangular blocks and we approximate the probability measure of the hidden Gibbs field by the product of some Gibbs distributions over the blocks. On that basis, we estimate the likelihood and derive the Block Likelihood Information Criterion (BLIC) that answers model choice questions such as the selection of the dependency structure or the number of latent states. We study the performances of BLIC for those questions. In addition, we present a comparison with ABC algorithms to point out that the novel criterion offers a better trade-off between time efficiency and reliable results.

show abstract

X^{n} = {({}, 1, \dots, K)}^{n}

parametrized by a scalar ψ that adjusts the level of dependence between adjacent sites and whose Hamiltonian is given by

H (x | ψ, G) = - ψ \sum_{iG \sim j} 1 {x_{i} = x_{j}} .

The aforementioned sum

iG \sim j

ranges the set of edges of the graph

G

.…”

Section: Comparison Of Bic Approximationsmentioning

confidence: 99%

Hidden Gibbs random fields model selection using Block Likelihood Information Criterion

Stoehr

Marin

Pudlo

2016

Stat

View full text Add to dashboard Cite

show abstract

“…Notable examples are the autologistic model (Besag, 1974) and its extension the Potts model. Shaped by the development of Geman and Geman (1984) and Besag (1986), -see for example Alfò et al (2008) and Moores et al (2014) who performed image segmentation with the help of this modelling -and also in other applications including disease mapping (e.g., Green and Richardson, 2002) and genetic analysis (e.g., François et al, 2006, Friel et al, 2009 to name a few. The exponential random graph model or p * model (Wasserman and Pattison, 1996) is another prominent example (Frank and Strauss, 1986) and arguably the most popular statistical model for social network analysis (e.g., Robins et al, 2007).…”

Section: Introductionmentioning

confidence: 99%

A review on statistical inference methods for discrete Markov random fields

Stoehr

2017

Preprint

View full text Add to dashboard Cite

Developing satisfactory methodology for the analysis of Markov random field is a very challenging task. Indeed, due to the Markovian dependence structure, the normalizing constant of the fields cannot be computed using standard analytical or numerical methods. This forms a central issue for any statistical approach as the likelihood is an integral part of the procedure. Furthermore, such unobserved fields cannot be integrated out and the likelihood evaluation becomes a doubly intractable problem. This report gives an overview of some of the methods used in the literature to analyse such observed or unobserved random fields.

show abstract

“…A potential automated solution for replicating delivered treatments was identified while conducting another study . Importing the OBI‐generated DICOM‐RT RS and RE files into Pinnacle 3 automatically creates and associates points of interest (POIs) with the CBCT (labelled AcqIsocentre, InitLaserIso and InitMatchIso) and registers the CBCT with the plan.…”

Section: Introductionmentioning

confidence: 99%

Automated replication of cone beam CT‐guided treatments in the Pinnacle³ treatment planning system for adaptive radiotherapy

Hargrave

Mason

Guidi

et al. 2015

J of Medical Radiation Sci

View full text Add to dashboard Cite

IntroductionTime‐consuming manual methods have been required to register cone‐beam computed tomography (CBCT) images with plans in the Pinnacle3 treatment planning system in order to replicate delivered treatments for adaptive radiotherapy. These methods rely on fiducial marker (FM) placement during CBCT acquisition or the image mid‐point to localise the image isocentre. A quality assurance study was conducted to validate an automated CBCT‐plan registration method utilising the Digital Imaging and Communications in Medicine (DICOM) Structure Set (RS) and Spatial Registration (RE) files created during online image‐guided radiotherapy (IGRT).Methods CBCTs of a phantom were acquired with FMs and predetermined setup errors using various online IGRT workflows. The CBCTs, DICOM RS and RE files were imported into Pinnacle3 plans of the phantom and the resulting automated CBCT‐plan registrations were compared to existing manual methods. A clinical protocol for the automated method was subsequently developed and tested retrospectively using CBCTs and plans for six bladder patients.ResultsThe automated CBCT‐plan registration method was successfully applied to thirty‐four phantom CBCT images acquired with an online 0 mm action level workflow. Ten CBCTs acquired with other IGRT workflows required manual workarounds. This was addressed during the development and testing of the clinical protocol using twenty‐eight patient CBCTs. The automated CBCT‐plan registrations were instantaneous, replicating delivered treatments in Pinnacle3 with errors of ±0.5 mm. These errors were comparable to mid‐point‐dependant manual registrations but superior to FM‐dependant manual registrations.ConclusionThe automated CBCT‐plan registration method quickly and reliably replicates delivered treatments in Pinnacle3 for adaptive radiotherapy.

show abstract

Segmentation of cone-beam CT using a hidden Markov random field with informative priors

Cited by 9 publications

References 8 publications

Hidden Gibbs random fields model selection using Block Likelihood Information Criterion

Hidden Gibbs random fields model selection using Block Likelihood Information Criterion

A review on statistical inference methods for discrete Markov random fields

Automated replication of cone beam CT‐guided treatments in the Pinnacle³ treatment planning system for adaptive radiotherapy

Contact Info

Product

Resources

About

Segmentation of cone-beam CT using a hidden Markov random field with informative priors

Cited by 9 publications

References 8 publications

Hidden Gibbs random fields model selection using Block Likelihood Information Criterion

Hidden Gibbs random fields model selection using Block Likelihood Information Criterion

A review on statistical inference methods for discrete Markov random fields

Automated replication of cone beam CT‐guided treatments in the Pinnacle3 treatment planning system for adaptive radiotherapy

Contact Info

Product

Resources

About

Automated replication of cone beam CT‐guided treatments in the Pinnacle³ treatment planning system for adaptive radiotherapy