Biomedical & Pharmacology Journal Vol. 7(2), 691-696 (2014) CT and MRI Brain Images Matching Using Ridgeness Correlation AYUSH DOGRA and PARVINDER BHALLA Departmenet of ECE, Maharishi Markandeshwar University, Mullana, Ambala, India. http://dx.doi.org/10.13005/bpj/543 (Received: August 10, 2014; accepted: November 05, 2014) ABSTRACT Image registration is considered as one of the most fundamental and crucial pre processing step in digital image processing. Since it is a vital problem in medical imaging, though it has several applications in clinical diagnosis such as diagnosis of cardiac, retinal, pelvic, renal, abdomen, liver, tissue etc disorders. Ridges, edges and troughs are useful geometric features for image analysis. Geometrical image features like ridges in digital images may be extracted by convolving the images with Gaussian kernels. In this paper we will perform the task of using ridgeness correlation for CT and MRI brain images matching using scale space method and also demonstrate 3D CT – MRI image registration with a software package using rigid translations and rotations. Key words: Ridges, Ridgeness, Troughs, Image registrations/matching, Multi resolution correlation INTRODUCTION N dimension are explained in2. Figure 1 shows clearly prominent ridge4. Image registration is a process of overlaying two images geometrically. Image In the application of CT and MRI registration can be classified into (1) multi view (2) registration, the image ridgeness seems to be useful multi temporal (3) multi modal. There are feature, as the skull ridge is most prominent in a CT abundance of image registration techniques. The image, and inverse ridge like – wise prominent in various areas where image registration is beneficial an MRI image. are remote sensing, image mosaicing, image fusion, medicine. The details of techniques can be Scale space representation found in1. The tomography imaging modalities such Scale space is representation of multi as CT, MRI, PET, SPECT are briefly explained in4,5,6. scale signals. It is developed by image processing While integrating the two modalities the first step is & signal processing communities. It is terminology image registration and second step is image to handle various image structures at different fusion16. In this paper ,we will discuss the first step scales. Scale space stated well in3,10. Scale space only. as stated by dr. petra van den elsen11,12. Ridges Measure of ridgeness Ridges are rough top of anything. It seems Many techniques have been developed to narrow elevation. Ridges are briefly explained to detect ridges over the century by Maxwell 1889, in2,4,7,8,9. Whereas troughs are reciprocal of ridge. koenderink & van doorn 1994. There are lot of The geometric definition of ridges & valleys in 2D & geometrical invariants to detect ridges in the image 692 DOGRA & BHALLA, Biomed. & Pharmacol. J., Vol. 7(2), 691-696 (2014) structures ( eberly at el. 1994). In this paper, selected space methods for detecting ridge. Figure 4 shows Lvv operator to detect ridge structures that is well the CT landscape but now smoothed by explained in11,13. convolution with a Gaussian. Now applying 3DLvv to CT set, the resultant image shows skull ridge. Why skull ridge? Now this ridged imaged is superimposed onto the Skull has deformable nature so it is original CT slice efficiently and precisely. Similar excellent structure of matching CT & MRI modalities examples can be shown in MR data sets (figure 5, of brian. Figure 2 shows CT slice, figure 3 depicts 6, 7) respectively. the landscape version of CT slice. This is basically an intensity landscape. In original image skull ridge is jagged (spikes). So resort to the use of scale Fig. 1: A Clearly Prominent Ridge Fig. 2: CT Slice Fig. 3: Intensity Landscape Version of Ct Slice Fig. 4: The Landscape Version of Previous in Previous Figure Figure Again Now Smoothed By Gaussian Convolution Fig. 5: Mri Data Set Fig. 6: Landscape Version of MR Data Set DOGRA & BHALLA, Biomed. & Pharmacol. J., Vol. 7(2), 691-696 (2014) 693 In MR image the skull area is dark so 3D Image registration of CT and MRI brain troughs or inverse ridge are detected. As we overlaid images[17] the ridgeness image onto the CT and MR datasets Till now, the matching of CT and MRI brain ,the 3DCT and MR ridgeness volumes(L1 and L2)is images was fully automatic. There was no time created as shown in figure(8&9) . consuming user interaction required and total users subjectivity was avoided. High computational effort Matching of 3D CT and MR Volumes was the disadvantage of above mentioned The registration of 3DCT and MR matching of CT and MRI images. Now we will ridgeness volume ( L1 and L2) using demonstrate the 3D CT and MRI registration where correlation5,11,12,13,14,15 of grey values, minimize c (t) the user interaction is needed. Thus it is called over rigid transformation, c (t) is defined as- interactive registration . With this interactive registration, a source image can be registered against a target image of different or same modality. In this method, user subjectivity is avoided, The source images shown in figure 12 and figure there is no user interaction & fully automatic. Only 13 shows target images. A source images is to be disadvantage related to this technique is high register on target images ,figure 14 is a registered computational effort required. image. RESULT This new registered image is formed and can be used for fusion and 3D modeling application. Resultant match of 3D CT ridgeness and The transforms (Rotation, Translation & Scaling) of 3D MR ridgeness is shown figure 10. Figure 11 source, target and the registered image is shown shows zoomed or detailed view of selected part. in table -1 (a, b, c) for source image, table -2 (d, e, f) for target image & table -3 (g, h, I) for registered image. Fig. 7: Landscape Version of Previous Figure Fig. 8: 3d CT Ridgeness Volume (L1) Now Blurred With Gaussian Kernel Fig. 9: 3D MR Ridgeness Volume (L2) Fig. 10: After Matching Using Ridgeness Correlation 694 DOGRA & BHALLA, Biomed. & Pharmacol. J., Vol. 7(2), 691-696 (2014) Fig. 11: Zoomed or Detailed View of Previous Fig. 12: MRI Source Image(courtesy-3D Image (Using Ridgeness Correlation) DOCTOR SOFTWARE) Fig. 13: CT Target Image(courtesy-3D DOCTOR Fig. 14: Registered Image (courtesy-3D SOFTWARE) DOCTOR SOFTWARE) Table 1: (a, b, c) (courtesy-3D DOCTOR Table 2: (d, e, f) (courtesy-3D DOCTOR SOFTWARE) SOFTWARE) Rotation Rotation X 0 X 0 Y 0 Y 0 Z 0 Z 0 (a) (d) Translation Translation X 0 X 0 Y 0 Y 0 Z 0 Z 0 (b) (e) Scale Scale X 1 X 1 Y 1 Y 1 Z 1 Z 1 (c) (f) DOGRA & BHALLA, Biomed. & Pharmacol. J., Vol. 7(2), 691-696 (2014) 695 Table 3: (g, h, I) (courtesy-3D DOCTOR Applications of CT-MRI registration SOFTWARE) Matching of MR & CT images of head can be useful in planning neuro surgical & ENT surgical Rotation procedures. The matching of CT & MRI modalities X 90 are used in radiotheraphy planning. It is used in Y 0 prophylactic cranial radiotherapy. Z 90 (g) CONCLUSION AND DISCUSSION Translation X -10 The terms matching and registration are Y -2 both used to donate the process of determining the Z -15 transformation that relates the content of two images in a meaning full way. In the former part of this paper, (h) we use 3D CT & MR ridgeness volumes in a multi Scale resolution correlation method. This scheme required X 1.1 no interactive actions and devoid of human Y 0.1 subjectivity. In the later part of this paper interactive Z 1.2 registration task is done where the user interaction (i) is needed. In these interactive registration patient related geometrical features is not needed. REFERENCES 1. B.Zitova and J.Flusser,”Image Registration 8. Eberly, D. (1996). Ridges in Image and Data methods:asurvey” Image and Vision Analysis. Kluwer. ISBN 0-7923-4268-2. Computing, pp.977 1000, (2003). 9. Kerrel, R. Generic Transitions of Relative 2. http://en.wikipedia.org/wiki/Ridge_detection Critical Sets in Parameterized Families with 3. http://en.wikipedia.org/wiki/Scale_space Applications to Image Analysis. University 4. Dogra, A., and M. S. Patterh. “CT and MRI of North Carolina. 1999. Brain Images Registration for Clinical 10. Graphical illustration of basic ideas of scale- Applications.” J Cancer Sci Ther 6: 018-026 space representation at www.csc.kth.se/ ((2014)). ~tony/cern- review/cern-html/node2.html 5. George T. Y. Chen and Charles A. Pelizzari. 11. J. B. A. Maintz, P. A. van den Elsen, and M. A Image Correlation Techniques in Radiation Viergever. Evaluation of ridge seeking Therapy Planning. Computerized Medical operators for multimodality medical image Imaging and Graphics, 13(3):235{240, 1989. matching. IEEE Trans. Pattern Anal. Mach. 6. C. F. Ru_, D. L. G. Hill, G. P. Robinson, and D. Intell., 18, pp. 353–365, Apr. 1996. J. Hawkes. Volume Rendering of Multimodal 12. J. B. A. Maintz, P. A. van den Elsen, and M. A. Images for the Planning of Skull Base Viergever, “Comparison of feature-based Surgery. In H. U. Lemke et al, editor, Computer matching of CT and MR brain images,” in Assisted Radiology ’93, pages 574 {579. Computer Vision, Virtual Reality, and Springer-Verlag, 1993. Robotics in Medicine 1995, N. Ayache, Ed. 7. Laptev, H. Mayer, T. Lindeberg, W. Eckstein, Berlin, Germany: Springer-Verlag, 1995, pp. C. Steger, A. Baumgartner. Automatic 219–228. extraction of roads from aerial images based 13. P. A. van den Elsen “Comparison of edge- on scale space and snakes. Machine Vision based and ridge-based registration of CT and Applications 12: 23–31, Springer-Verlag and MR brain images,” Med. Image Anal., 1, 2000 pp. 151–161, (1996). 696 DOGRA & BHALLA, Biomed. & Pharmacol. J., Vol. 7(2), 691-696 (2014) 14. C. Studholme, D. L. G. Hill, and D. J. Hawkes, _MRI _Brain_Images “Automated registration of truncated MR and 16. Ayush at el.”An efficient data level fusion of CT datasets of the head,” Proc. Br. Mach. multimodal medical images by cross scale Vision Assoc., pp. 27–36 (1995). fusion rule”. International Journal of 15. http://www.academia.edu/6849310/ Advanced Scientific and Technical Global_Journal_of_Medical_research_ Research, Issue 4 volume 5, Sep. – Oct. 2014 Neurology_and_Nervous_System_ 17. http://www.ablesw.com/3d-doctor/regist.html Feature_ based _Matching _of _CT _and