DocumentCode :
3242692
Title :
Automated segmentation of the thyroid gland on CT using multi-atlas label fusion and random forest
Author :
Jiamin Liu ; Narayanan, Divya ; Chang, Kevin ; Kim, Lauren ; Turkbey, Evrim ; Le Lu ; Jianhua Yao ; Summers, Ronald M.
Author_Institution :
Imaging Biomarkers & Comput.-aided Diagnosis Lab., Nat. Inst. of Health Clinical Center, Bethesda, MD, USA
fYear :
2015
fDate :
16-19 April 2015
Firstpage :
1114
Lastpage :
1117
Abstract :
The thyroid gland is an important endocrine organ. For a variety of clinical applications, a system for automated segmentation of the thyroid is desirable. Thyroid segmentation is challenging due to the inhomogeneous nature of the thyroid and the surrounding structures which have similar intensities. In this paper, we propose a fully automated method for thyroid detection and segmentation on CT scans. The thyroid gland is initially estimated by a multi-atlas segmentation with joint label fusion algorithm. The segmentation is then corrected by supervised statistical learning-based voxel labeling with a random forest algorithm. Multi-atlas label fusion transfers expert-labeled thyroids from atlases to a target image using deformable registration. Errors produced by label transfer are reduced by label fusion that combines the results produced by all atlases into a consensus solution. Then, random forest employs an ensemble of decision trees that are trained on labeled thyroids to recognize various features. The trained forest classifier is then applied to the estimated thyroid by voxel scanning to assign the class-conditional probability. Voxels from the expert-labeled thyroids in CT volumes are treated as positive classes and background non-thyroid voxels as negatives. We applied our method to 73 patients using 5 as atlases. The system achieved an overall 0.70 Dice Similarity Coefficient (DSC) if using the multi-atlas label fusion only and was improved to 0.75 DSC after the random forest correction.
Keywords :
biological organs; computerised tomography; decision trees; feature extraction; image classification; image fusion; image reconstruction; image registration; image segmentation; learning (artificial intelligence); medical image processing; random processes; statistical analysis; CT scans; automated thyroid gland segmentation; class-conditional probability; decision trees; deformable registration; dice similarity coefficient; endocrine organ; feature recognition; joint label fusion algorithm; multiatlas label fusion; random forest algorithm; supervised statistical learning-based voxel labeling; Computed tomography; Glands; Image segmentation; Joints; Labeling; Training; Vegetation; multiatlas label fusion; random forest; thyroid gland segmentation;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Biomedical Imaging (ISBI), 2015 IEEE 12th International Symposium on
Conference_Location :
New York, NY
Type :
conf
DOI :
10.1109/ISBI.2015.7164067
Filename :
7164067
Link To Document :
بازگشت