近幾年醫療器材迅速蓬勃發展，其中高階影像相關器材占六成以上的醫療器材市場。2017年4月中華民國經濟部(Ministry of Economic Affairs, R.O.C.)投入約20億元於高階影像醫療器材，其中「高階即時影像」和「多功能整合型監控」更是不可或缺的部分。因此醫療用電腦輔助診斷系統(Computer aided detection/diagnosis, CAD)運用影像註冊技術將病人拍攝的醫療影像進行重建及結合，提供醫師更完整的醫療資訊、降低病灶的診斷和判讀時間成為高階影像醫療器材的重要一環。醫學影像註冊技術除了協助醫師進行疾病判讀，也可進行病患疾病的監測和分析，現今更為了降低微創手術系統(Minimally invasive surgery, MIS)的困難度，將其與手術導航技術結合(Surgical navigation systems)，並運用手術導航系統中的影像註冊技術進行空間定位及影像對位，建構可視化的整合資訊，提供醫師完善且簡易的操作。

由於脊椎導航系統從2012年8.7億美元預估增長至2018年的14.5億美元左右，脊椎影像註冊儼然成為醫學影像註冊中研究重點之一，因此本論文主要使用3D CT 和 MRI 脊椎影像透過建立機器學習LSVM模型進行脊椎節偵測，利用新開發的VLSA方法(Vertebra localization signal analysis method)自動找尋特徵點，並將特徵點結合三種不同的影像註冊方式產生3D CT 和 3D MRI註冊結果，並將其提供整合性脊椎資訊，協助醫師在手術前後的觀察和判讀，進而提升醫療品質和技術。

本研究開發一全自動3D CT 和 MRI 脊椎影像註冊系統，可將兩種3D立體醫療影像進行立體影像融和。於技術驗證上，使用手動標註15個相對結構點於 CT 和 MRI 影像上進行準確度量化分析。由於先前實驗中已使用手動標註15個相對結構點於比較九種註冊方法，因此我們挑選前兩組(分別為 FASGD_MI_Si_Bs 和 FASGD_MI_Af_Bs)速度較快且準確度較高的演算法進行註冊結果的比較。經由驗證結果比較三種本研究提出的三種演算法及上述兩種演算法進行註冊，本研究達到最小平均誤差(<6mm)，相較於FASGD_MI_Si_Bs的平均誤差(>16mm)及FASGD_Mi_Af_Bs的平均誤差(>20mm)，本研究註冊結果可提供醫師較為準確的註冊影像。此外，本研究也對使用醫師手動標點及本研究的全自動找點方式進行比較，實驗結果透過成對樣本T檢定獲得p>0.05，證明本研究全自動找尋特徵點與醫生手動標點有相似的結果。此技術的設計期望可以協助醫師進生行病患診斷和監控，降低醫師的判斷時間，同時也期盼能與醫療器材進行整合，開發出更完善的醫療設備，促進整體的醫療品質。

This paper presents an fully automatic registration and fusion system of 3D CT and 3D MRI datasets of the spine regions. The automatic system is consisted of a spine detection method, a landmark detection approach, a corresponding landmark detection model and an elastic 4D registration approach.

In evaluation, a preliminary test has been conducted to compare nine registration methods with the presented registration approaches using five manually identified corresponding landmarks, and the top two benchmark methods with high registration accuracies and computing speed are selected as the benchmark methods for full evaluation.Next, using the outputs of the proposed automatic corresponding landmark detection approach, we compare the proposed three registration methods with the selected top two benchmark methods to identify the optimal 4D alignment method. Then, we compare the performance of the same registration model using manually selected corresponding landmarks versus using our automatic landmark detection results.

Full evaluation utilizes fifteen manually similar anatomic features on CT and MRI spine images to calculate the average distance error for qualitative comparative analysis.Specifically, for the benchmark method with 3D CT and MR datasets of the spine regions, the first datasets achieved for a mean distance error of 12.9128 pixels(<6mm) and for second datasets a mean distance error of 5.7344 pixels(<6mm). With use of a two-tailed Student t test for paired samples in the comparing the fully automatic registration and semi-automatic registration. For the both datasets there were no significant difference in the automatic registration when compared with a semi-automatic registration(where p > 0.05).

The results show that we presented registration method perform the proposed method is significantly better than top two benchmark methods (p $\leq$ 0.001). In addition, the results show that the registration accuracy of the registration method using the automatic detected corresponding landmarks is similar to the method using the manually identified landmarks.

[1] D. L. Hill, P. G. Batchelor, M. Holden, D. J. Hawkes, Medical image registration, Physics in medicine and biology 46 (3) (2001) R1.
[2] B. Zitova, J. Flusser, Image registration methods: a survey, Image and vision
computing 21 (11) (2003) 977–1000.
[3] C. Studholme, D. L. Hill, D. J. Hawkes, Automated 3-d registration of mr and
ct images of the head, Medical image analysis 1 (2) (1996) 163–175.
[4] M. Jenkinson, P. Bannister, M. Brady, S. Smith, Improved optimization for the
robust and accurate linear registration and motion correction of brain images,
Neuroimage 17 (2) (2002) 825–841.
[5] C. Studholme, D. L. Hill, D. J. Hawkes, An overlap invariant entropy measure
of 3d medical image alignment, Pattern recognition 32 (1) (1999) 71–86.
[6] H. Lester, S. R. Arridge, A survey of hierarchical non-linear medical image
registration, Pattern recognition 32 (1) (1999) 129–149.
[7] Y. Otake, A. S. Wang, J. W. Stayman, A. Uneri, G. Kleinszig, S. Vogt, A. J.
Khanna, Z. L. Gokaslan, J. H. Siewerdsen, Robust 3d–2d image registration:
application to spine interventions and vertebral labeling in the presence of
anatomical deformation, Physics in medicine and biology 58 (23) (2013) 8535.
[8] M. R. Safari, P. Rowe, A. McFadyen, A. Buis, Hands-off and hands-on casting
consistency of amputee below knee sockets using magnetic resonance imaging,
The Scientific World Journal 2013.
[9] X. Dong, M. A. Gonzalez Ballester, G. Zheng, Automatic extraction of femur
contours from calibrated x-ray images using statistical information., Journal of
multimedia 2 (5).
[10] K. Rohr, H. S. Stiehl, R. Sprengel, T. M. Buzug, J. Weese, M. Kuhn, Landmarkbased elastic registration using approximating thin-plate splines, IEEE Transactions on medical imaging 20 (6) (2001) 526–534.
[11] M. Ershad, A. Ahmadian, N. D. Serej, H. Saberi, K. A. Khoiy, Minimization
of target registration error for vertebra in image-guided spine surgery, International journal of computer assisted radiology and surgery 9 (1) (2014) 29–38.
[12] C. Maurer, R. J. Maciunas, J. M. Fitzpatrick, Registration of head ct images
to physical space using a weighted combination of points and surfaces [imageguided surgery], IEEE transactions on medical imaging 17 (5) (1998) 753–761.
[13] D. Andreasen, K. Van Leemput, R. H. Hansen, J. A. Andersen, J. M. Edmund,
Patch-based generation of a pseudo ct from conventional mri sequences for mrionly radiotherapy of the brain, Medical physics 42 (4) (2015) 1596–1605.
[14] N. Burgos, M. J. Cardoso, K. Thielemans, M. Modat, S. Pedemonte, J. Dickson, A. Barnes, R. Ahmed, C. J. Mahoney, J. M. Schott, et al., Attenuation
correction synthesis for hybrid pet-mr scanners: application to brain studies,
IEEE transactions on medical imaging 33 (12) (2014) 2332–2341.
[15] S. Klein, M. Staring, K. Murphy, M. A. Viergever, J. P. Pluim, Elastix: a
toolbox for intensity-based medical image registration, IEEE transactions on
medical imaging 29 (1) (2010) 196–205.
[16] S. Roy, A. Carass, A. Jog, J. L. Prince, J. Lee, Mr to ct registration of brains
using image synthesis, in: Proceedings of SPIE, Vol. 9034, NIH Public Access,
2014.
[17] C. Dean, J. Sykes, R. Cooper, P. Hatfield, B. Carey, S. Swift, S. Bacon,
D. Thwaites, D. Sebag-Montefiore, A. Morgan, An evaluation of four ct–mri
co-registration techniques for radiotherapy treatment planning of prone rectal
cancer patients, The British journal of radiology 85 (1009) (2012) 61–68.
[18] S. Gill, P. Abolmaesumi, G. Fichtinger, J. Boisvert, D. Pichora, D. Borshneck,
P. Mousavi, Biomechanically constrained groupwise ultrasound to ct registration of the lumbar spine, Medical Image Analysis 16 (3) (2012) 662–674.
[19] R. H. Gong, O. G¨uler, M. K¨urkl¨uoglu, J. Lovejoy, Z. Yaniv, Interactive initial- ¨
ization of 2d/3d rigid registration, Medical physics 40 (12).
[20] I. Hacihaliloglu, A. Rasoulian, R. N. Rohling, P. Abolmaesumi, Local phase
tensor features for 3-d ultrasound to statistical shape+ pose spine model registration, IEEE transactions on Medical Imaging 33 (11) (2014) 2167–2179.
[21] R. Harmouche, F. Cheriet, H. Labelle, J. Dansereau, 3d registration of mr and
x-ray spine images using an articulated model, Computerized Medical Imaging
and Graphics 36 (5) (2012) 410–418.
[22] J. H. Jonsson, P. Brynolfsson, A. Garpebring, M. Karlsson, K. S¨oderstr¨om,
T. Nyholm, Registration accuracy for mr images of the prostate using a subvolume based registration protocol, Radiation Oncology 6 (1) (2011) 73.
[23] A. Lang, P. Mousavi, S. Gill, G. Fichtinger, P. Abolmaesumi, Multi-modal
registration of speckle-tracked freehand 3d ultrasound to ct in the lumbar spine,
Medical image analysis 16 (3) (2012) 675–686.
[24] S. Nagpal, P. Abolmaesumi, A. Rasoulian, I. Hacihaliloglu, T. Ungi, J. Osborn,
V. A. Lessoway, J. Rudan, M. Jaeger, R. N. Rohling, et al., A multi-vertebrae
ct to us registration of the lumbar spine in clinical data, International journal
of computer assisted radiology and surgery 10 (9) (2015) 1371–1381.
[25] D. Rivest-H´enault, N. Dowson, P. B. Greer, J. Fripp, J. A. Dowling, Robust inverse-consistent affine ct–mr registration in mri-assisted and mri-alone
prostate radiation therapy, Medical image analysis 23 (1) (2015) 56–69.
[26] W. Wein, S. Brunke, A. Khamene, M. R. Callstrom, N. Navab, Automatic ctultrasound registration for diagnostic imaging and image-guided intervention,
Medical image analysis 12 (5) (2008) 577–585.
[27] J. Shi, W. Zhang, M. Tang, R. J. Caselli, Y. Wang, Conformal invariants for
multiply connected surfaces: Application to landmark curve-based brain morphometry analysis, Medical image analysis 35 (2017) 517–529.
[28] A. Seginer, Rigid-body point-based registration: The distribution of the target
registration error when the fiducial registration errors are given, Medical image
analysis 15 (4) (2011) 397–413.
[29] D. Zhi, Limitations of improving neuronavigation registration accuracy by adjusting the fiducial point with the largest fiducial registration error, Journal of
Medical Imaging and Health Informatics 6 (7) (2016) 1692–1695.
[30] E. C. Chen, A. J. McLeod, J. S. Baxter, T. M. Peters, Registration of 3d shapes
under anisotropic scaling, International journal of computer assisted radiology
and surgery 10 (6) (2015) 867–878.
[31] Y. Otake, M. Armand, R. S. Armiger, M. D. Kutzer, E. Basafa, P. Kazanzides,
R. H. Taylor, Intraoperative image-based multiview 2d/3d registration for
image-guided orthopaedic surgery: incorporation of fiducial-based c-arm tracking and gpu-acceleration, IEEE transactions on medical imaging 31 (4) (2012)
948–962.
[32] A. Varnavas, T. Carrell, G. Penney, Increasing the automation of a 2d-3d registration system, IEEE transactions on medical imaging 32 (2) (2013) 387–399.
[33] H. Chen, C. Shen, J. Qin, D. Ni, L. Shi, J. C. Cheng, P.-A. Heng, Automatic
localization and identification of vertebrae in spine ct via a joint learning model
with deep neural networks, in: International Conference on Medical Image
Computing and Computer-Assisted Intervention, Springer, 2015, pp. 515–522.
[34] M. Lootus, T. Kadir, A. Zisserman, Vertebrae detection and labelling in lumbar mr images, in: Computational methods and clinical applications for spine
imaging, Springer, 2014, pp. 219–230.
[35] A. B. Oktay, Y. S. Akgul, Localization of the lumbar discs using machine learning and exact probabilistic inference, in: International Conference on Medical
Image Computing and Computer-Assisted Intervention, Springer, 2011, pp.
158–165.
[36] A. Raja’S, J. J. Corso, V. Chaudhary, Labeling of lumbar discs using both
pixel-and object-level features with a two-level probabilistic model, IEEE transactions on medical imaging 30 (1) (2011) 1–10.
[37] A. Suzani, A. Seitel, Y. Liu, S. Fels, R. N. Rohling, P. Abolmaesumi, Fast
automatic vertebrae detection and localization in pathological ct scans-a deep
learning approach, in: International Conference on Medical Image Computing
and Computer-Assisted Intervention, Springer, 2015, pp. 678–686.
[38] Y. Zhan, D. Maneesh, M. Harder, X. S. Zhou, Robust mr spine detection using
hierarchical learning and local articulated model, in: International Conference
on Medical Image Computing and Computer-Assisted Intervention, Springer,
2012, pp. 141–148.
[39] G. Zheng, C. Chu, D. L. Belav`y, B. Ibragimov, R. Korez, T. Vrtovec, H. Hutt,
R. Everson, J. Meakin, I. L. Andrade, et al., Evaluation and comparison of 3d
intervertebral disc localization and segmentation methods for 3d t2 mr data: A
grand challenge, Medical image analysis 35 (2017) 327–344.
[40] P. Felzenszwalb, D. McAllester, D. Ramanan, A discriminatively trained, multiscale, deformable part model, in: Computer Vision and Pattern Recognition,
2008. CVPR 2008. IEEE Conference on, IEEE, 2008, pp. 1–8.
[41] C. Kalavagunta, X. Zhou, S. C. Schmechel, G. J. Metzger, Registration of in
vivo prostate mri and pseudo-whole mount histology using local affine transformations guided by internal structures (latis), Journal of Magnetic Resonance
Imaging 41 (4) (2015) 1104–1114.
[42] S. Gu, X. Meng, F. C. Sciurba, H. Ma, J. Leader, N. Kaminski, D. Gur,
J. Pu, Bidirectional elastic image registration using b-spline affine transformation, Computerized Medical Imaging and Graphics 38 (4) (2014) 306–314.
[43] J.-P. Thirion, Image matching as a diffusion process: an analogy with maxwell’s
demons, Medical image analysis 2 (3) (1998) 243–260.
[44] R. Szeliski, J. Coughlan, Spline-based image registration, International Journal
of Computer Vision 22 (3) (1997) 199–218.
[45] T. Kanai, N. Kadoya, K. Ito, Y. Onozato, S. Y. Cho, K. Kishi, S. Dobashi,
R. Umezawa, H. Matsushita, K. Takeda, et al., Evaluation of accuracy of bspline transformation-based deformable image registration with different parameter settings for thoracic images, Journal of radiation research 55 (6) (2014)
1163–1170.
[46] S. Srinivasan, B. A. Hargreaves, B. L. Daniel, Fat-based registration of breast
dynamic contrast enhanced water images, Magnetic Resonance in Medicine.
[47] Y. Qiao, B. van Lew, B. P. Lelieveldt, M. Staring, Fast automatic step size
estimation for gradient descent optimization of image registration, IEEE transactions on medical imaging 35 (2) (2016) 391–403.
[48] Y. Qiao, B. P. Lelieveldt, M. Staring, Fast automatic estimation of the optimization step size for nonrigid image registration., in: Medical Imaging: Image
Processing, 2014, p. 90341A.
[49] A. Hammers, R. Allom, M. J. Koepp, S. L. Free, R. Myers, L. Lemieux, T. N.
Mitchell, D. J. Brooks, J. S. Duncan, Three-dimensional maximum probability
atlas of the human brain, with particular reference to the temporal lobe, Human
brain mapping 19 (4) (2003) 224–247.
[50] D. Rueckert, L. I. Sonoda, C. Hayes, D. L. Hill, M. O. Leach, D. J. Hawkes,
Nonrigid registration using free-form deformations: application to breast mr
images, IEEE transactions on medical imaging 18 (8) (1999) 712–721.
[51] J. West, J. M. Fitzpatrick, M. Y. Wang, B. M. Dawant, C. R. Maurer Jr, R. M.
Kessler, R. J. Maciunas, C. Barillot, D. Lemoine, A. Collignon, et al., Comparison and evaluation of retrospective intermodality brain image registration
techniques, Journal of computer assisted tomography 21 (4) (1997) 554–568.
[52] J. Stolk, H. Putter, E. M. Bakker, S. B. Shaker, D. G. Parr, E. Piitulainen,
E. W. Russi, E. Grebski, A. Dirksen, R. A. Stockley, et al., Progression parameters for emphysema: a clinical investigation, Respiratory medicine 101 (9)
(2007) 1924–1930.
[53] K. Murphy, B. van Ginneken, S. Klein, M. Staring, B. J. de Hoop, M. A.
Viergever, J. P. Pluim, Semi-automatic construction of reference standards for
evaluation of image registration, Medical Image Analysis 15 (1) (2011) 71–84.
[54] M. Staring, M. Bakker, J. Stolk, D. Shamonin, J. Reiber, B. Stoel, Towards
local progression estimation of pulmonary emphysema using ct, Medical physics
41 (2).
[55] P. F. Felzenszwalb, R. B. Girshick, D. McAllester, D. Ramanan, Object detection with discriminatively trained part-based models, IEEE transactions on
pattern analysis and machine intelligence 32 (9) (2010) 1627–1645.
[56] N. Dalal, B. Triggs, Histograms of oriented gradients for human detection, in:
Computer Vision and Pattern Recognition, 2005. CVPR 2005. IEEE Computer
Society Conference on, Vol. 1, IEEE, 2005, pp. 886–893.
[57] B. Li, K. Cheng, Z. Yu, Histogram of oriented gradient based gist feature for
building recognition, Computational intelligence and neuroscience 2016 (2016)
10.
[58] A. B. Oktay, Y. S. Akgul, Simultaneous localization of lumbar vertebrae and
intervertebral discs with svm-based mrf, IEEE Transactions on Biomedical Engineering 60 (9) (2013) 2375–2383.
[59] B. Akay, A study on particle swarm optimization and artificial bee colony algorithms for multilevel thresholding, Applied Soft Computing 13 (6) (2013)
3066–3091.
[60] J. N. Kapur, P. K. Sahoo, A. K. Wong, A new method for gray-level picture
thresholding using the entropy of the histogram, Computer vision, graphics,
and image processing 29 (3) (1985) 273–285.
[61] T. Huang, S. Blostein, E. Margerum, Least-squares estimation of motion parameters from 3-d point correspondences, in: Proc. IEEE Conf. Computer Vision and Pattern Recognition, Vol. 10, 1986, pp. 112–115.
[62] B. K. Horn, Closed-form solution of absolute orientation using unit quaternions,
JOSA A 4 (4) (1987) 629–642.
[63] P. Th´evenaz, M. Unser, Optimization of mutual information for multiresolution
image registration, IEEE transactions on image processing 9 (12) (2000) 2083–
2099.