Segmentation propagation from deformable atlases for brain mapping and analysis

TY - CHAP

T1 - Segmentation propagation from deformable atlases for brain mapping and analysis

AU - Linguraru, Marius George

AU - Vercauteren, Tom

AU - Reyes-Aguirre, Mauricio

AU - Ballester, Miguel Ángel González

AU - Ayache, Nicholas

PY - 2011/1/1

Y1 - 2011/1/1

N2 - Magnetic resonance imaging (MRI) is commonly employed for the depiction of softtissues, most notably the human brain. Computer-aided image analysis techniques lead toimage enhancement and automatic detection of anatomical structures. However, theintensity information contained in images does not often offer enough contrast to robustlyobtain a good detection of all internal brain structures, not least the deep gray matternuclei. We propose digital atlases that deform to fit the image data to be analyzed. In thisapplication, deformable atlases are employed for the detection and segmentation of brainnuclei, to allow analysis of brain structures. Our fully automatic technique is based on acombination of rigid, affine and non-linear registration, a priori information on keyanatomical landmarks and propagation of the information of the atlas. The Internet BrainSegmentation Repository (IBSR) data provide manually segmented brain data. Usingprior anatomical knowledge in local brain areas from a randomly chosen brain scan(atlas), a first estimation of the deformation fields is calculated by affine registration. Theimage alignment is refined through a non-linear transformation to correct thesegmentation of nuclei. The local segmentation results are greatly improved. They arerobust over the patient data and in accordance with the clinical ground truth. Validationof results is assessed by comparing the automatic segmentation of deep gray nuclei by theproposed method with manual segmentation. The technique offers the accuratesegmentation of difficultly identifiable brain structures in conjuncture with deformableatlases. Such automated processes allow the study of large image databases and provideconsistent measurements over the data. The method has a wide range of clinicalapplications of high impact that span from size and intensity quantification tocomprehensive (anatomical, functional, dynamic) analysis of internal brain structures.

AB - Magnetic resonance imaging (MRI) is commonly employed for the depiction of softtissues, most notably the human brain. Computer-aided image analysis techniques lead toimage enhancement and automatic detection of anatomical structures. However, theintensity information contained in images does not often offer enough contrast to robustlyobtain a good detection of all internal brain structures, not least the deep gray matternuclei. We propose digital atlases that deform to fit the image data to be analyzed. In thisapplication, deformable atlases are employed for the detection and segmentation of brainnuclei, to allow analysis of brain structures. Our fully automatic technique is based on acombination of rigid, affine and non-linear registration, a priori information on keyanatomical landmarks and propagation of the information of the atlas. The Internet BrainSegmentation Repository (IBSR) data provide manually segmented brain data. Usingprior anatomical knowledge in local brain areas from a randomly chosen brain scan(atlas), a first estimation of the deformation fields is calculated by affine registration. Theimage alignment is refined through a non-linear transformation to correct thesegmentation of nuclei. The local segmentation results are greatly improved. They arerobust over the patient data and in accordance with the clinical ground truth. Validationof results is assessed by comparing the automatic segmentation of deep gray nuclei by theproposed method with manual segmentation. The technique offers the accuratesegmentation of difficultly identifiable brain structures in conjuncture with deformableatlases. Such automated processes allow the study of large image databases and provideconsistent measurements over the data. The method has a wide range of clinicalapplications of high impact that span from size and intensity quantification tocomprehensive (anatomical, functional, dynamic) analysis of internal brain structures.

KW - Atlas

KW - Brain

KW - Deformation

KW - Gray matter nuclei

KW - MRI

KW - Registration

KW - Segmentation

UR - http://www.scopus.com/inward/record.url?scp=84892131820&partnerID=8YFLogxK

M3 - Chapter

AN - SCOPUS:84892131820

SN - 9781611220650

SP - 179

EP - 196

BT - Brain Mapping and Diseases

PB - Nova Science Publishers Inc

ER -