Vertebral Shape: Automatic Measurement with Active Shape Models

Smyth, Paul; Taylor, Christopher J.; Adams, Judith E.

doi:10.1148/radiology.211.2.r99ma40571

Cited by 98 publications

(77 citation statements)

References 17 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The images are assigned a score for each mode describing how many standard deviations they lie from the mean of all the images. The model may also be used to locate and characterise similar objects in new images (11)(12). The purpose of the current study was to create a shape model of the lumbar spine in the sagittal plane to characterize the lordosis and to compare the reliability and precision of the model with that of conventional lordosis measurements.…”

mentioning

confidence: 99%

See 1 more Smart Citation

Characterizing the Shape of the Lumbar Spine Using an Active Shape Model

Meakin¹,

Gregory²,

Smith³

et al. 2008

Spine

View full text Add to dashboard Cite

Acknowledgements:We thank Dr Y Hirasawa (MD) and Mrs B MacLennan (research radiographer) for acquiring the MR images.Modelling the lumbar spine shape 2 Structured Abstract Study DesignAnalysis of positional magnetic resonance images of normal volunteers. ObjectivesTo compare the reliability and precision of an active shape model to that of conventional lordosis measurements/ Summary of Background DataCharacterisation of lumbar lordosis commonly relies on measurement of angles; these have been found to have errors of around 10 o . MethodsT2 weighted sagittal images of the lumbar spines of 24 male volunteers in the standing posture were acquired using a positional MR scanner. An active shape model of the vertebral bodies from S1 to L1 was created. Lumbar lordosis was also determined by measuring the angles of the superior end-plates. All measurements were performed twice by one observer and once by a second observer. ResultsThe shape model identified two modes of variation to describe the shape of the lumbar spine (mode 1 described curvature and mode 2 described evenness of curvature). Significant correlations were found between mode 1 and total lordosis (R = 0.97, P < 0.001) and between mode 2 and mean absolute deviation of segmental lordosis (R = 0.80, P < 0.001). Intra-and inter-observer reliability was higher for the shape model (ICCs 0.98 -1.00) than for the Modelling the lumbar spine shape 3 lordosis angle measurements (ICCs 0.68 -0.99). The relative error of the shape model (mode 1 = 4 %; mode 2 = 9 %) was lower than the conventional measurements (total lordosis = 10 %). ConclusionsThe shape of the lumbar spine in the sagittal plane can be comprehensively characterised using a shape model. The results are more reliable and precise than measurements of lordosis calculated from end-plate angles. Key wordsLumbar spine; lordosis; positional MRI; Active Shape Model Mini AbstractAn active shape model was used to characterize the shape of the lumbar spine from positional MRI images of 24 male volunteers in the standing posture. The results of the model were more reliable and precise than the conventional method that uses end-plate angle measurements to characterize the lumbar lordosis.Modelling the lumbar spine shape 4

show abstract

mentioning

confidence: 99%

“…Although the main focus of our study was to characterise shape, a trained active shape model may be used to automatically locate an object in an image (11)(12). This would be particularly beneficial for analysing a time series of images, such as obtained using…”

mentioning

confidence: 99%

Characterizing the Shape of the Lumbar Spine Using an Active Shape Model

Meakin¹,

Gregory²,

Smith³

et al. 2008

Spine

View full text Add to dashboard Cite

show abstract

“…The approach is broadly applicable since, for example, segmentation of overlapping objects is almost inevitable in medical applications (e.g. [11]). Depending on the actual application, different likelihood functions might have to be adopted for example to account for differing imaging modalities.…”

Section: Discussionmentioning

confidence: 99%

Segmenting Multiple Objects with Overlapping Appearance and Uncertainty

Seise

McKenna

Ricketts

et al. 2006

Procedings of the British Machine Vision Conference 2006

View full text Add to dashboard Cite

A probabilistic method is proposed for segmentation of multiple objects that overlap or are in close proximity to one another. A likelihood function is formulated that explicitly models overlapping object appearance. Priors on global appearance and geometry (including shape) are learned from example images. Markov chain Monte Carlo methods are used to obtain samples from a posterior distribution over model parameters from which expectations can be estimated. The method is described in detail for the problem of segmenting femur and tibia in x-ray images. The result is a probabilistic segmentation that quantifies uncertainty so that measurements such as joint space can be made with associated uncertainty.

show abstract

“…Zamora et al reported success rates of 50% with average errors below 6.4 mm in active shape model (ASM) segmentation of lumbar vertebrae in spine radiographs [20]. Smyth et al performed ASM segmentation of vertebrae in dual energy X ray absorptiometry (DXA) images [16] and obtained success rates of 94 -98%, with errors in the order of 1 mm for healthy vertebra and success rates of 85 -98% with errors in the order of 2 mm for fractured vertebrae. Scott et al reported successful convergence of a modified active appearance model (AAM) in 92% of DXA scans of healthy spines with an average error of ca.…”

Section: Discussionmentioning

confidence: 99%

“…Linear models of object appearance were shown to fail in many medical image segmentation tasks [18,9,4,14]. Another drawback of current deformable model approaches it that they require initialization near the final solution, and thus need manual intervention [9,16] or automatic object recognition [5,20].…”

Section: Introductionmentioning

confidence: 99%

Shape Particle Filtering for Image Segmentation

Bruijne

Nielsen

2004

Medical Image Computing and Computer-Assisted Intervention – MICCAI 2004

View full text Add to dashboard Cite

Abstract. Deformable template models are valuable tools in medical image segmentation. Current methods elegantly incorporate global shape and appearance, but can not cope with localized appearance variations and rely on an assumption of Gaussian gray value distribution. Furthermore, initialization near the optimal solution is required. We propose a maximum likelihood shape inference that is based on pixel classification, so that local and non-linear intensity variations are dealt with naturally, while a global shape model ensures a consistent segmentation. Optimization by stochastic sampling removes the need for accurate initialization. The method is demonstrated on three different medical image segmentation problems: vertebra segmentation in spine radiographs, lung field segmentation in thorax X rays, and delineation of the myocardium of the left ventricle in MRI slices. Accurate results were obtained in all tasks.

show abstract

Vertebral Shape: Automatic Measurement with Active Shape Models

Cited by 98 publications

References 17 publications

Characterizing the Shape of the Lumbar Spine Using an Active Shape Model

Characterizing the Shape of the Lumbar Spine Using an Active Shape Model

Segmenting Multiple Objects with Overlapping Appearance and Uncertainty

Shape Particle Filtering for Image Segmentation

Contact Info

Product

Resources

About