2017
DOI: 10.1155/2017/8612519
|View full text |Cite
|
Sign up to set email alerts
|

Head and Neck Cancer Tumor Segmentation Using Support Vector Machine in Dynamic Contrast-Enhanced MRI

Abstract: Objective We aimed to propose an automatic method based on Support Vector Machine (SVM) and Dynamic Contrast-Enhanced Magnetic Resonance Imaging (DCE-MRI) to segment the tumor lesions of head and neck cancer (HNC). Materials and Methods 120 DCE-MRI samples were collected. Five curve features and two principal components of the normalized time-intensity curve (TIC) in 80 samples were calculated as the dataset in training three SVM classifiers. The other 40 samples were used as the testing dataset. The area over… Show more

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
2
1
1
1

Citation Types

1
7
0

Year Published

2019
2019
2023
2023

Publication Types

Select...
7
2

Relationship

1
8

Authors

Journals

citations
Cited by 22 publications
(8 citation statements)
references
References 9 publications
1
7
0
Order By: Relevance
“…In similar studies, authors obtained a HR of 86% and MR of 0.79, which was considered satisfactory. (24)(25)(26) In this study, the HR was 89.23% and the MR was 0.70, i.e., quantitatively similar to those of previous authors, which indicates good accuracy of the proposed segmentation method. The best HR results were obtained for T2 and FSPGR-T1c sequences.…”
Section: Comparative Performance Analysis Of Automatic Segmentationsupporting
confidence: 89%
See 1 more Smart Citation
“…In similar studies, authors obtained a HR of 86% and MR of 0.79, which was considered satisfactory. (24)(25)(26) In this study, the HR was 89.23% and the MR was 0.70, i.e., quantitatively similar to those of previous authors, which indicates good accuracy of the proposed segmentation method. The best HR results were obtained for T2 and FSPGR-T1c sequences.…”
Section: Comparative Performance Analysis Of Automatic Segmentationsupporting
confidence: 89%
“…Quantitative results of the automatic segmentation were calculated by comparison between the gold standard and the proposed segmentation method. The metrics commonly used in the literature (24)(25)(26) to assess performance are the hit rate (HR), calculated by the number of true positive (TP) of the method compared with the gold standard; and the matching rate (MR), based on the number of false positive results (FP). In the cited studies, the HR and MR are respectively defined as:…”
Section: Evaluation Of Resultsmentioning
confidence: 99%
“…However, it is hard to formulate an objective, robust and straightforward cost function for graph-based methods due to potentially contradicting / conflicting requirements among image. Recently, machine learning based segmentation approaches, such as Support Vector Machine (Deng et al, 2017), Decision tree (Berthon et al, 2017) and k-nearest neighbour (KNN) (Yu et al, 2009;Comelli et al, 2018), are adopted in many head and neck segmentation studies. These classifiers make decisions for each voxel by gradients or texture features extracted from the neighbourhood without any shape constraint.…”
Section: Introductionmentioning
confidence: 99%
“…They trained on a publicly available data set of 250 patients and were able to achieve a Dice coefficient of 0.73. In another study, Deng et al 57 developed a tumor segmentation algorithm with a training data set of 120 contrast-enhanced head and neck MRI scans. The model was developed using support vector machines (a method of classical ML) and was able to achieve an area overlap measure of 0.76.…”
Section: Head and Neckmentioning
confidence: 99%