A Bayesian model of shape and appearance for subcortical brain segmentation
By Brian Patenaude, Stephen M. Smith, Daniel P. Kennedy, Mark Jenkinson. NeuroImage, 2011
2,344 citations recorded by the source corpus.
Automatic segmentation of subcortical structures in human brain MR images is an important but difficult task due to poor and variable intensity contrast. Clear, well-defined intensity features are absent in many places along typical structure boundaries and so extra information is required to achieve successful segmentation. A method is proposed here that uses manually labelled image data to provide anatomical training information. It utilises the principles of the Active Shape and Appearance Models but places them within a Bayesian framework, allowing probabilistic relationships between shape and intensity to be fully exploited. The model is trained for 15 different subcortical structures using 336 manually-labelled T1-weighted MR images. Using the Bayesian approach, conditional probabilities can be calculated easily and efficiently, avoiding technical problems of ill-conditioned covariance matrices, even with weak priors, and eliminating the need for fitting extra empirical scaling parameters, as is required in standard Active Appearance Models. Furthermore, differences in boundary vertex locations provide a direct, purely local measure of geometric change in structure between groups that, unlike voxel-based morphometry, is not dependent on tissue classification methods or arbitrary smoothing. In this paper the fully-automated segmentation method is presented and assessed both quantitatively, using Leave-One-Out testing on the 336 training images, and qualitatively, using an independent clinical dataset involving Alzheimer’s disease. Median Dice overlaps between 0.7 and 0.9 are obtained with this method, which is comparable or better than other automated methods. An implementation of this method, called FIRST, is currently distributed with the freely-available FSL package.
Subfields: Computer Vision and Pattern Recognition, Geometry and Topology
View Paper: https://www.semanticscholar.org/paper/ece35ca3c934fce37b357231fd1dfe26f65e8c1d
How researchers see this paper
The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.
In Computer Vision and Pattern Recognition
- Bruce Fischl
- Adrian V. Dalca
- José Vicente Manjón
- Stephen M. Smith
- John V. Guttag
- Guha Balakrishnan
- Douglas N. Greve
- Pierrick Coupé
In Geometry and Topology
- Andreas Horn
- Andrea A. Kühn
- José Vicente Manjón
- Pierrick Coupé
- Arthur W. Toga
- Douglas N. Greve
- Randy L. Buckner
- Lars Tjelta Westlye
Citations in the corpus
- Lead-DBS: A toolbox for deep brain stimulation electrode localizations and visualizations (2015, positive): “So far, however, most current subcortical segmentation algorithms do not or only partly support classical DBS target regions (Fischl et al., 2002; Patenaude et al., 2011).”
- Image processing and Quality Control for the first 10,000 brain imaging datasets from UK Biobank (2017, background): “Volumes of different tissue types and total brain volume, both unnormalised and normalised for head (2011)).”
- Anatomically-constrained tractography: Improved diffusion MRI streamlines tractography through effective use of anatomical information (2012, methodology): “In this context, FIRST was configured to automatically segment the nucleus accumbens, caudate, globus pallidus, putamen and thalamus (all bilaterally).”
- volBrain: An Online MRI Brain Volumetry System (2015, methodology): “One of these tools, called FIRST (Patenaude et al., 2011), is able to automatically segment subcortical brain structures.”
- The ENIGMA Consortium: large-scale collaborative analyses of neuroimaging and genetic data (2014, positive): “Nugent et al. (2012) also studied the inter-scanner reliability of the FIRST software for segmentation (Patenaude et al. 2011), and found it to be high.”
- Sex differences in the adult human brain: Evidence from 5,216 UK Biobank participants (2017, methodology): “Subcortical segmentation was also conducted by the UK Biobank imaging team using FIRST (Patenaude et al. 2011) to provide the volumes of 15 structures (atlas in Fig.”
- Multimodal population brain imaging in the UK Biobank prospective epidemiological study (2016, neutral): “While this might not provide statistical certainty for a diagnosis or interventional recommendation, it should allow for single-patient imaging to be used in a similar way to current state-of-the-art patient-tuned genetic testing.”
- 3D fully convolutional networks for subcortical segmentation in MRI: A large‐scale study (2016, methodology): “[61] (2011) Bayesian model Thalamus Caudate Putamen Pallidum 0.”
- Data Augmentation Using Learned Transformations for One-Shot Medical Image Segmentation (2019, methodology): “AAMs have been used to localize anatomical landmarks [17, 58] and perform segmentation [52, 56, 77].”
- Deep learning based brain tumor segmentation: a survey (2020, background): “This segmentation assumption is that the brain image does not contain any tumor tissue or other anomalies [7], [8].”
- Performing Label-Fusion-Based Segmentation Using Multiple Automatically Generated Templates (2012, methodology): “Our results when using three input templates for MAGeT Brain are also comparable to those demonstrated in the recent release of the FIRST algorithm [Patenaude et al., 2011] which reports Kappas of (cid:2) 0.898, 0.871, 0.821, and 0.894 for the caudate, putamen, globus pallidus, and thalamus…”
- Rationale, design and methodology of the image analysis protocol for studies of patients with cerebral small vessel disease and mild stroke (2015, methodology): “uk/fsl/fslwiki/ FIRST) (Patenaude et al. 2011) optimized for the current imaging protocol and incorporating boundary correction in the hippocampus and in the caudate nucleus.”
- A probabilistic atlas of the human thalamic nuclei combining ex vivo MRI and histology (2018, methodology): “Patenaude et al. (2011) used a combined model of shape and appearance, also to segment a set of brain structures including the thalamus; an implementation of this method (“FIRST”) is available as part of the popular FSL package (Smith et al., 2004).”