A machine learning approach to extract spinal column centerline from three-dimensional CT data

Wang, Caihua; Li, Yuanzhong; Ito, Wataru; Shimura, Kazuo; Abe, Katsumi

doi:10.1117/12.810982

Cited by 9 publications

(10 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Centerline detection on medical images based on offline boosting has been reported elsewhere [19] with some successful results. The boosting is a linear combination of weak classifiers called stumps , which can be any classifier performing slightly better than random guessing such as a one-layer decision tree [19].…”

Section: Methodsmentioning

confidence: 99%

See 1 more Smart Citation

Robust infrarenal aortic aneurysm lumen centerline detection for rupture status classification

Zhang

Kheyfets

Finol

2013

Medical Engineering & Physics

View full text Add to dashboard Cite

The objective of this work is to develop a robust method for human abdominal aortic aneurysm (AAA) centerline detection that can contribute to the accurate computation of features for the prediction of AAA rupture risk. A semiautomatic algorithm is proposed for detecting the lumen centerline in contrast-enhanced abdominal computed tomography images based on online adaboost classifiers, which does not require prior image segmentation. The algorithm was developed and applied to thirty ruptured and thirty unruptured AAA image data sets and the tortuosities of the detected centerline were measured to assess the correlation between AAA tortuosity and the binary ruptured and unruptured labels. The lumen of each data set was segmented manually by a trained radiologist and the resulting centerlines of each data set were defined as the gold standard to evaluate the accuracy of the algorithm and to compare it against two widely used segmentation techniques. The average mean relative accuracy of the offline adaboost classifier is 91.9% with a standard deviation of 1.6%; for the online adaboost classifier it is 93.6% with a standard deviation of 1.9% (p<0.05). The online adaboost classifier outperforms the offline adaboost classifier while their computational costs are similar. Aneurysm tortuosity computed from an accurately derived lumen centerline using online adaboost is statistically higher for ruptured aneurysms compared to unruptured aneurysms, indicating that tortuosity can be used to assess rupture risk in the vascular clinic.

show abstract

Section: Methodsmentioning

confidence: 99%

“…The boosting is a linear combination of weak classifiers called stumps , which can be any classifier performing slightly better than random guessing such as a one-layer decision tree [19]. During the training step, all training data are assigned weights equally, and at each iteration a weak classifier, f m ( ), is applied on the training data.…”

Section: Methodsmentioning

confidence: 99%

Robust infrarenal aortic aneurysm lumen centerline detection for rupture status classification

Zhang

Kheyfets

Finol

2013

Medical Engineering & Physics

View full text Add to dashboard Cite

show abstract

“…Other techniques, such as [7], [20], [56], [60], aim at extracting the centerlines as we do, but still rely on binary classification to distinguish the image locations on centerlines from the rest. [7], [60] use Haar wavelets in conjunction with boosted trees to detect the centerlines of tubular structures at different scales.…”

Section: Learned Filtersmentioning

confidence: 99%

“…[20] uses spectral-structural features instead and SVMs to find road centerlines. In [56] co-occurrence features and the AdaBoost algorithm are used to detect the spinal column centerline.…”

Section: Learned Filtersmentioning

confidence: 99%

Multiscale Centerline Detection

Sironi

Türetken

Lepetit

et al. 2016

IEEE Trans. Pattern Anal. Mach. Intell.

113

View full text Add to dashboard Cite

Abstract-Finding the centerline and estimating the radius of linear structures is a critical first step in many applications, ranging from road delineation in 2D aerial images to modeling blood vessels, lung bronchi, and dendritic arbors in 3D biomedical image stacks. Existing techniques rely either on filters designed to respond to ideal cylindrical structures or on classification techniques. The former tend to become unreliable when the linear structures are very irregular while the latter often has difficulties distinguishing centerline locations from neighboring ones, thus losing accuracy. We solve this problem by reformulating centerline detection in terms of a regression problem. We first train regressors to return the distances to the closest centerline in scale-space, and we apply them to the input images or volumes. The centerlines and the corresponding scale then correspond to the regressors local maxima, which can be easily identified. We show that our method outperforms state-of-the-art techniques for various 2D and 3D datasets. Moreover, our approach is very generic and also performs well on contour detection. We show an improvement above recent contour detection algorithms on the BSDS500 dataset.

show abstract

“…Three landmarks are detected by applying AdaBoost detectors on multi-planar reconstruction images (Fig. 2 (Right)) [14]. After detecting the landmarks from each data, the 3D coordinates are normalized by shift, scale, and rotation.…”

Section: Shape Modelmentioning

confidence: 99%

Automatic coronary extraction by supervised detection and shape matching

Kitamura

Ito

2012

2012 9th IEEE International Symposium on Biomedical Imaging (ISBI)

Self Cite

View full text Add to dashboard Cite

Automatic coronary extraction has great clinical importance in the effective handling and visualization of large amounts of 3D data. Despite tremendous previous research, coronary extraction remains difficult. Two such difficulties are extraction of both normal and abnormal vessels and reconstruction of exact tree structures based on anatomical knowledge.To solve the first difficulty, we propose a method to learn a classifier of a tubular 3D object with a dimension reduction approach using Hessian analysis. This enables detection of vessel candidate points despite variations in their appearances. Regarding the second difficulty, we propose an approach to apply the MRF framework for vascular structure segmentation. A novelty of the approach is incorporating constraints to avoid topological inconsistency. Correspondences between the candidate points and model points are found using a graph matching process during which, tree structures as per the shape model are simultaneously reconstructed. Experimental results show robustness of the method. The proposed method can improve clinical workflow.

show abstract

A machine learning approach to extract spinal column centerline from three-dimensional CT data

Cited by 9 publications

References 0 publications

Robust infrarenal aortic aneurysm lumen centerline detection for rupture status classification

Robust infrarenal aortic aneurysm lumen centerline detection for rupture status classification

Multiscale Centerline Detection

Automatic coronary extraction by supervised detection and shape matching

Contact Info

Product

Resources

About