3D temporal subtraction on chest MDCT images using nonlinear image registration technique

Transcription

1 3D temporal subtraction on chest MDCT images using nonlinear image registration technique Poster No.: C-1072 Congress: ECR 2010 Type: Topic: Authors: Keywords: DOI: Scientific Exhibit Computer Applications P. Yan, Y. Kodera; Nagoya/JP cheat CT image, 3D nonlinear image warping, 3D temporal subtraction /ecr2010/C-1072 Any information contained in this pdf file is automatically generated from digital material submitted to EPOS by third parties in the form of scientific presentations. References to any names, marks, products, or services of third parties or hypertext links to thirdparty sites or information are provided solely as a convenience to you and do not in any way constitute or imply ECR's endorsement, sponsorship or recommendation of the third party, information, product or service. ECR is not responsible for the content of these pages and does not make any representations regarding the content or accuracy of material in this file. As per copyright regulations, any unauthorised use of the material or parts thereof as well as commercial reproduction or multiple distribution by any traditional or electronically based reproduction/publication method ist strictly prohibited. You agree to defend, indemnify, and hold ECR harmless from and against any and all claims, damages, costs, and expenses, including attorneys' fees, arising from or related to your use of these pages. Please note: Links to movies, ppt slideshows and any other multimedia files are not available in the pdf version of presentations. Page 1 of 9

2 Purpose The purpose is to perform the 3D temporal subtraction algorithm for chest MDCT images by using 3D image registration method, in order to enhance the interval changes in lung nodules. Methods and Materials Image data We used chest CT images that were obtained by a MDCT scanner (Aquilion16; TOSHIBA Medical Systems, JP) in Nagoya University Hospital. The original CT images are 512(X) 512(Y) slices(z) in matrix size and 0.64mm(X) 0.64mm(Y) 5mm(Z) in voxel size. Overall Scheme The scheme for 3D temporal subtraction comprises four steps: 1. Normalize voxel size For maintaining the continuity of image data along the Z-axis, the voxel size of the previous and current images were normalized by using the linear interpolation method. Then, thickness of the image data was converted from 5mm to 1.0mm. Page 2 of 9

3 2. Global image registration After the voxel size adjustment, the smooth processing on previous and current images performed respectively to decrease the noise. The rectangular region of interest (ROI) of lung section image was selected from each current slice. The 2D normalized crosscorrelation coefficient was used to register each ROI region to the previous image. The corresponding image at the best-matched location was determined when the crosscorrelation value was at its maximum between a current image and a previous image, then the best matched previous was extracted, as shown in Fig Local image registration In this part, we selected template areas from the current image and the corresponding search areas from the previous image. The dimensions of the template areas were voxels and of the search areas were voxels. respectively. Fig.2 shows the template areas of size voxels obtained from the current image. Fig.3 shows the search areas of size voxels obtained from the previous image. Then, we used a normalized cross-correlation coefficient to calculate the both areas for executing a local registration. The shift values of #x, #y and #z were then obtained from the maximum value of the cross-correlation coefficient. The previous image was then arranged according to the fitted shift values. 4.Image subtraction The subtraction image was obtained by subtracting the current image from the arranged previous one. Page 3 of 9

4 Images for this section: Fig. 1: Fig.1 Page 4 of 9

5 Fig. 2: Fig.2 Page 5 of 9

6 Fig. 3: Fig.3 Page 6 of 9

7 Results The normal background structures such as vessels and heat are removed by using 3D temporal subtraction.thus,the changes in lungs that occurred from diseases are clearly enhanced as shadows in the subtraction images. Fig.4 shows previous images and currenet images with lung model. Fig.5 shows subtraction images without and with the image registration. Images for this section: Page 7 of 9

8 Fig. 1: Fig.4 Fig. 2: Fig.5 Page 8 of 9

9 Conclusion We performed the 3D temporal subtraction algorithm based on the global and local matching using 3D image registration method, the 3D temporal subtraction image would be useful for enhance interval changes on MDCT images. References 1. Ishida T, Katsuragawa S, Kawashita I, Kim H, Itai Y, Awai K, Q LI, Doi K. 3D temporal subtraction on multislice CT images using nonlinear warping technique. 2007; Proc SPIE 6414: Ishida T, Katsuragawa S, Nakamura K, MacMahon H, Doi K. Iterative image warping technique for temporal subtraction of sequential chest radiographs to detect interval change Med Phys 26 (7): Peng W, Andrea D, Paul O, Watler GO. Lung metastases detection in CT images using 3D template matching Med Phys 34 (3): Kano A, Doi K, MacMahon H, Hassell DD, Maryellen LG. Digital image subtraction of temporally sequential chest images for detection of interval change Med Phys 21 (3): Personal Information Page 9 of 9