JZUS - Journal of Zhejiang University SCIENCE

Frontiers of Information Technology & Electronic Engineering 2018 Vol.19 No.4 P.471-480

http://doi.org/10.1631/FITEE.1620342

Kernel sparse representation for MRI image analysis in automatic brain tumor segmentation

Author(s): Ji-jun Tong, Peng Zhang, Yu-xiang Weng, Dan-hua Zhu
Affiliation(s): School of Information Science and Technology, Zhejiang Sci-Tech University, Hangzhou 310018, China; more
Corresponding email(s): jijuntong@zstu.edu.cn
Key Words: Brain tumor segmentation, Kernel method, Sparse coding, Dictionary learning

Share this article to： More \|Next Article >>>

Abstract: The segmentation of brain tumor plays an important role in diagnosis, treatment planning, and surgical simulation. The precise segmentation of brain tumor can help clinicians obtain its location, size, and shape information. We propose a fully automatic brain tumor segmentation method based on kernel sparse coding. It is validated with 3D multiple-modality magnetic resonance imaging (MRI). In this method, MRI images are pre-processed first to reduce the noise, and then kernel dictionary learning is used to extract the nonlinear features to construct five adaptive dictionaries for healthy tissues, necrosis, edema, non-enhancing tumor, and enhancing tumor tissues. sparse coding is performed on the feature vectors extracted from the original MRI images, which are a patch of m×m×m around the voxel. A kernel-clustering algorithm based on dictionary learning is developed to code the voxels. In the end, morphological filtering is used to fill in the area among multiple connected components to improve the segmentation quality. To assess the segmentation performance, the segmentation results are uploaded to the online evaluation system where the evaluation metrics dice score, positive predictive value (PPV), sensitivity, and kappa are used. The results demonstrate that the proposed method has good performance on the complete tumor region (dice: 0.83; PPV: 0.84; sensitivity: 0.82), while slightly worse performance on the tumor core (dice: 0.69; PPV: 0.76; sensitivity: 0.80) and enhancing tumor (dice: 0.58; PPV: 0.60; sensitivity: 0.65). It is competitive to the other groups in the brain tumor segmentation challenge. Therefore, it is a potential method in differentiation of healthy and pathological tissues.

基于核稀疏表示的磁共振图像分析及其在脑肿瘤自动分割中的应用

摘要：脑肿瘤分割在疾病辅助诊断、治疗方案规划以及手术导航中扮演重要角色。对脑肿瘤精确分割可以帮助临床医生获取肿瘤位置、尺寸和形状信息。提出一种基于核稀疏编码的全自动脑肿瘤分割方法，并在3D多模态磁共振成像图（magnetic resonance imaging, MRI）上验证。首先对MRI图像进行预处理以减少噪声，然后通过核字典学习提取非线性特征，用来构建坏死组织、水肿组织、非增强肿瘤组织、增强肿瘤组织和健康组织5个适应性字典。对从原始MRI图像上肿瘤像素点周边m×m×m的小区域提取的特征向量进行稀疏编码，并通过一种基于字典学习的核聚类方法对像素点进行编码。最后通过形态滤波填充在多个相连部分间的区域，提高分割质量。为评估分割表现，分割结果被上传到在线评估系统中，该评估系统使用dice系数、阳性预测值（positive predictive value, PPV）、灵敏度和kappa值作为评估指标。结果表明，该方法在完整肿瘤区域分割上具有良好表现（dice: 0.83; PPV: 0.84; sensitivity: 0.82），而在肿瘤核心区域（dice: 0.69; PPV: 0.76; sensitivity: 0.80）和增强肿瘤区域（dice: 0.58; PPV: 0.60; sensitivity: 0.65）上表现稍差。相较于脑肿瘤分割（BRATS）挑战中其他团队采用的方法，该方法具有竞争力。该方法在健康组织和病理组织区分上具有一定潜力。

关键词：脑肿瘤分割；核方法；稀疏编码；字典学习

Reference

[1]Ahmadvand A, Daliri MR, 2015. Improving the runtime of MRF based method for MRI brain segmentation. Appl Math Comput, 256:808-818.

[2]Atkins MS, Mackiewich BT, 1998. Fully automatic segmentation of the brain in MRI. IEEE Trans Med Imag, 17(1):98-107.

[3]Bryt O, Elad M, 2008. Compression of facial images using the K-SVD algorithm. J Vis Commun Imag Represent, 19(4): 270-282.

[4]Chen SS, Donoho DL, Saunders MA, 2001. Atomic decomposition by basis pursuit. SIAM Rev, 43(1):129-159.

[5]Chong VFH, Zhou JY, Khoo JBK, et al., 2004. Tongue carcinoma: tumor volume measurement. Int J Radiat Oncol Biol Phys, 59(1):59-66.

[6]Cristianini N, Shawe-Taylor J, 2000. An Introduction to Support Vector Machines and Other Kernel-Based Learning Methods. Cambridge University Press, Cambridge, p.189.

[7]Dong WS, Zhang L, Shi GM, 2011. Centralized sparse representation for image restoration. Proc IEEE Int Conf on Computer Vision, p.1259-1266.

[8]Duarte-Carvajalino JM, Sapiro G, 2009. Learning to sense sparse signals: simultaneous sensing matrix and sparsifying dictionary optimization. IEEE Trans Image Process, 18(7):1395-1408.

[9]Dvořák P, Menze B, 2015. Structured prediction with convolutional neural networks for multimodal brain tumor segmentation. Proc Multimodal Brain Tumor Image Segmentation Challenge, p.13-24.

[10]Elad M, Aharon M, 2006a. Image denoising via learned dictionaries and sparse representation. Proc IEEE Computer Society Conf on Computer Vision and Pattern Recognition, p.895-900.

[11]Elad M, Aharon M, 2006b. Image denoising via sparse and redundant representations over learned dictionaries. IEEE Trans Image Process, 15(12):3736-3745.

[12]Fletcher-Heath LM, Hall LO, Goldgof DB, et al., 2001. Automatic segmentation of non-enhancing brain tumors in magnetic resonance images. Artif Intell Med, 21(1-3): 43-63.

[13]Gibbs P, Buckley DL, Blackband SJ, et al., 1996. Tumour volume determination from MR images by morphological segmentation. Phys Med Biol, 41(11):2437-2446.

[14]Grosse R, Raina R, Kwong H, et al., 2012. Shift-invariance sparse coding for audio classification. http://arxiv.org/abs/1206.5241

[15]He ZS, Cichocki A, Li YQ, et al., 2009. K-hyperline clustering learning for sparse component analysis. Signal Process, 89(6):1011-1022.

[16]Held K, Kops ER, Krause BJ, et al., 1997. Markov random field segmentation of brain MR images. IEEE Trans Med Imag, 16(6):878-886.

[17]Hyvärinen A, Hoyer P, Oja E, 1999. Image denoising by sparse code shrinkage. Proc Intelligent Signal Processing, p.1-31.

[18]Juan-Albarracin J, Fuster-Garcia E, Manjon JV, et al., 2015. Automated glioblastoma segmentation based on a multiparametric structured unsupervised classification. PLoS ONE, 10(5):e0125143.

[19]Juergens KU, Seifarth H, Range F, et al., 2008. Automated threshold-based 3D segmentation versus short-axis planimetry for assessment of global left ventricular function with dual-source MDCT. Am J Roentgenol, 190(2): 308-314.

[20]Kistler M, Bonaretti S, Pfahrer M, et al., 2013. The virtual skeleton database: an open access repository for biomedical research and collaboration. J Med Int Res, 15(11): e245.

[21]Kong YY, Li YJ, Wu JS, et al., 2016. Noise reduction of diffusion tensor images by sparse representation and dictionary learning. BioMed Eng, 15:5.

[22]Liu J, Li M, Wang JX, et al., 2014. A survey of MRI-based brain tumor segmentation methods. Tsinghua Sci Technol, 19(6):578-595.

[23]Mairal J, Elad M, Sapiro G, 2008. Sparse representation for color image restoration. IEEE Trans Imag Process, 17(1):53-69.

[24]Mairal J, Bach F, Ponce J, et al., 2009. Non-local sparse models for image restoration. Proc 12^th Int Conf on Computer Vision, p.2272-2279.

[25]Menze BH, Jakab A, Bauer S, et al., 2015. The multimodal brain tumor image segmentation benchmark (BRATS). IEEE Trans Med Imag, 34(10):1993-2024.

[26]Mittelhäußer G, Kruggel F, 1995. Fast segmentation of brain magnetic resonance tomograms. Proc 1^st Int Conf on Computer Vision, Virtual Reality and Robotics in Medicine, p.237-241.

[27]Nasir M, Baig A, Khanum A, 2014. Brain tumor classification in MRI scans using sparse representation. In: Elmoataz A, Lezoray O, Nouboud F, et al. (Eds.), Image and Signal Processing. Springer, Cham, p.629-637.

[28]Olabarriaga SD, Smeulders AWM, 2001. Interaction in the segmentation of medical images: a survey. Med Imag Anal, 5(2):127-142.

[29]Prastawa M, Bullitt E, Ho S, et al., 2004. A brain tumor segmentation framework based on outlier detection. Med Imag Anal, 8(3):275-283.

[30]Rathi VPGP, Palani S, 2015. Brain tumor detection and classification using deep learning classifier on MRI images. Res J Appl Sci Eng Technol, 10(2):177-187.

[31]Rousson M, Lenglet C, Deriche R, et al., 2004. Level set and region based surface propagation for diffusion tensor MRI segmentation. In: Sonka M, Kakadiaris IA, Kybic J (Eds.), Computer Vision and Mathematical Methods in Medical and Biomedical Image Analysis. Springer Berlin Heidelberg, p.123-134.

[32]Ruan S, Bloyet D, 2000. MRF models and multifractal analysis for MRI segmentation. Proc 5^th Int Conf on Signal Processing, p.1259-1262.

[33]Sachdeva J, Kumar V, Gupta I, et al., 2013. Segmentation, feature extraction, and multiclass brain tumor classification. J Dig Imag, 26(6):1141-1150.

[34]Salman Al-Shaikhli SD, Yang MY, Rosenhahn B, 2015. Brain tumor classification and segmentation using sparse coding and dictionary learning. BioMed Tech (Berl), 61(4): 413-429.

[35]Salman YM, Assal MA, Badawi AM, et al., 2005. Validation techniques for quantitative brain tumors measurements. Proc 27^th Annual Int Conf of the Engineering in Medicine and Biology Society, p.7048-7051.

[36]Shanthi KJ, Kumar MS, 2007. Skull stripping and automatic segmentation of brain MRI using seed growth and threshold techniques. Proc Int Conf on Intelligent and Advanced Systems, p.422-426.

[37]Sivaram GSVS, Nemala SK, Elhilali M, et al., 2010. Sparse coding for speech recognition. Proc IEEE Int Conf on Acoustics, Speech and Signal Processing, p.4346-4349.

[38]Sompong C, Wongthanavasu S, 2014. MRI brain tumor segmentation using GLCM cellular automata-based texture feature. Proc Int Computer Science and Engineering Conf, p.192-197.

[39]Taheri S, Ong SH, Chong VFH, 2010. Level-set segmentation of brain tumors using a threshold-based speed function. Imag Vis Comput, 28(1):26-37.

[40]Thiagarajan JJ, Ramamurthy KN, Spanias A, 2011. Optimality and stability of the K-hyperline clustering algorithm. Patt Recogn Lett, 32(9):1299-1304.

[41]Thiagarajan JJ, Ramamurthy KN, Rajan D, et al., 2014. Kernel sparse models for automated tumor segmentation. Int J Artif Intell Tools, 23(3):1460004.

[42]Tong T, Wolz R, Coupé P, et al., 2013. Segmentation of MR images via discriminative dictionary learning and sparse coding: application to hippocampus labeling. NeuroImage, 76:11-23.

[43]Tustison N, Wintermark M, Durst C, et al., 2013. ANTs and arboles. Proc NCI-MICCAI BRATS, p.47-50.

[44]Wang ZZ, Vemuri BC, 2004. Tensor field segmentation using region based active contour model. In: Pajdla T, Matas J (Eds.), Computer Vision-ECCV. Springer Berlin Heidelberg, p.304-315.

[45]Wong K, 2005. Medical image segmentation: methods and applications in functional imaging. In: Suri JS, Wilson DL, Laxminarayan S (Eds.), Handbook of Biomedical Image Analysis, Volume II: Segmentation Models Part B. Springer, Boston, US, p.111-182.

[46]Wu P, Xie K, Zheng Y, et al., 2012. Brain tumors classification based on 3D shape. In: Jin D, Lin S (Eds.), Advances in Future Computer and Control Systems. Springer, Berlin, p.277-283.

[47]Yang JC, Yu K, Gong YH, et al., 2009. Linear spatial pyramid matching using sparse coding for image classification. Proc IEEE Conf on Computer Vision and Pattern Recognition, p.1794-1801.

[48]Zeyde R, Elad M, Protter M, et al., 2012. On single image scale-up using sparse-representations. In: Boissonnat J, Chenin P, Cohen A, et al. (Eds.), Curves and Surfaces. Springer, Berlin, p.711-730.

Open peer comments: Debate/Discuss/Question/Opinion

<1>

Similar articles

- Go to

基于核稀疏表示的磁共振图像分析及其在脑肿瘤自动分割中的应用

Darkslateblue:Affiliate; Royal Blue:Author; Turquoise:Article

Reference