Limits...
Mjolnir: extending HAMMER using a diffusion transformation model and histogram equalization for deformable image registration.

Ellingsen LM, Prince JL - Int J Biomed Imaging (2009)

Bottom Line: The method, called Mjolnir, is an extension of the highly successful method HAMMER.New image features in order to better localize points of correspondence between the two images are introduced as well as a novel approach to generate a dense displacement field based upon the weighted diffusion of automatically derived feature correspondences.The results were compared with results generated by HAMMER and are shown to yield significant improvements in cortical alignment as well as reduced computation time.

View Article: PubMed Central - PubMed

Affiliation: Department of Electrical and Computer Engineering, The Johns Hopkins University, Baltimore, MD 21218, USA. lotta@jhu.edu

ABSTRACT
Image registration is a crucial step in many medical image analysis procedures such as image fusion, surgical planning, segmentation and labeling, and shape comparison in population or longitudinal studies. A new approach to volumetric intersubject deformable image registration is presented. The method, called Mjolnir, is an extension of the highly successful method HAMMER. New image features in order to better localize points of correspondence between the two images are introduced as well as a novel approach to generate a dense displacement field based upon the weighted diffusion of automatically derived feature correspondences. An extensive validation of the algorithm was performed on T1-weighted SPGR MR brain images from the NIREP evaluation database. The results were compared with results generated by HAMMER and are shown to yield significant improvements in cortical alignment as well as reduced computation time.

No MeSH data available.


Related in: MedlinePlus

Demonstration of the effects of histogram equalization on the segmentation. The top row shows the original image without histogram equalization on the left and a histogram equalized image on the right. The bottom row shows the corresponding hard segmented images. The arrows show smoother edges and better connected regions of the CSF class when histogram equalization is applied before segmentation. Arbitrarily located CSF spots have been removed, thereby increasing the consistency of the CSF class between different subjects.
© Copyright Policy - open-access
Related In: Results  -  Collection


getmorefigures.php?uid=PMC2724857&req=5

fig3: Demonstration of the effects of histogram equalization on the segmentation. The top row shows the original image without histogram equalization on the left and a histogram equalized image on the right. The bottom row shows the corresponding hard segmented images. The arrows show smoother edges and better connected regions of the CSF class when histogram equalization is applied before segmentation. Arbitrarily located CSF spots have been removed, thereby increasing the consistency of the CSF class between different subjects.

Mentions: In Mjolnir, we use histogram equalization [33] of both images in order to create a more exact match between the intensities of the two images and to create sulcal features that are typically quite subtle in the original MR images. This procedure reduces differences and inconsistencies between different subject scans as well as increasing the intensity contrast between tissue classes, thereby improving their ability to define distinctive landmarks. In particular, we have noticed that histogram equalization yields hard segmentations that are more revealing of tissue class boundaries that are consistently defined across different subjects. This property is illustrated in Figure 3. In this example, regions of sulcal CSF are arbitrarily disconnected in the hard segmentation of the original image, whereas the CSF within sulci is largely connected when the histogram equalized image is used. This tends to increase the consistency of the CSF class between different subjects. A specific example on improved consistency between different subjects is shown in Figure 4.


Mjolnir: extending HAMMER using a diffusion transformation model and histogram equalization for deformable image registration.

Ellingsen LM, Prince JL - Int J Biomed Imaging (2009)

Demonstration of the effects of histogram equalization on the segmentation. The top row shows the original image without histogram equalization on the left and a histogram equalized image on the right. The bottom row shows the corresponding hard segmented images. The arrows show smoother edges and better connected regions of the CSF class when histogram equalization is applied before segmentation. Arbitrarily located CSF spots have been removed, thereby increasing the consistency of the CSF class between different subjects.
© Copyright Policy - open-access
Related In: Results  -  Collection

Show All Figures
getmorefigures.php?uid=PMC2724857&req=5

fig3: Demonstration of the effects of histogram equalization on the segmentation. The top row shows the original image without histogram equalization on the left and a histogram equalized image on the right. The bottom row shows the corresponding hard segmented images. The arrows show smoother edges and better connected regions of the CSF class when histogram equalization is applied before segmentation. Arbitrarily located CSF spots have been removed, thereby increasing the consistency of the CSF class between different subjects.
Mentions: In Mjolnir, we use histogram equalization [33] of both images in order to create a more exact match between the intensities of the two images and to create sulcal features that are typically quite subtle in the original MR images. This procedure reduces differences and inconsistencies between different subject scans as well as increasing the intensity contrast between tissue classes, thereby improving their ability to define distinctive landmarks. In particular, we have noticed that histogram equalization yields hard segmentations that are more revealing of tissue class boundaries that are consistently defined across different subjects. This property is illustrated in Figure 3. In this example, regions of sulcal CSF are arbitrarily disconnected in the hard segmentation of the original image, whereas the CSF within sulci is largely connected when the histogram equalized image is used. This tends to increase the consistency of the CSF class between different subjects. A specific example on improved consistency between different subjects is shown in Figure 4.

Bottom Line: The method, called Mjolnir, is an extension of the highly successful method HAMMER.New image features in order to better localize points of correspondence between the two images are introduced as well as a novel approach to generate a dense displacement field based upon the weighted diffusion of automatically derived feature correspondences.The results were compared with results generated by HAMMER and are shown to yield significant improvements in cortical alignment as well as reduced computation time.

View Article: PubMed Central - PubMed

Affiliation: Department of Electrical and Computer Engineering, The Johns Hopkins University, Baltimore, MD 21218, USA. lotta@jhu.edu

ABSTRACT
Image registration is a crucial step in many medical image analysis procedures such as image fusion, surgical planning, segmentation and labeling, and shape comparison in population or longitudinal studies. A new approach to volumetric intersubject deformable image registration is presented. The method, called Mjolnir, is an extension of the highly successful method HAMMER. New image features in order to better localize points of correspondence between the two images are introduced as well as a novel approach to generate a dense displacement field based upon the weighted diffusion of automatically derived feature correspondences. An extensive validation of the algorithm was performed on T1-weighted SPGR MR brain images from the NIREP evaluation database. The results were compared with results generated by HAMMER and are shown to yield significant improvements in cortical alignment as well as reduced computation time.

No MeSH data available.


Related in: MedlinePlus