RISS 검색 - 국내학술지논문

1
Seeds Classification for Image Segmentation Based on 3-D Affine Moment Invariants

Anwar Abdalbari,Jing Ren,Mark Green 대한의용생체공학회 2016 Biomedical Engineering Letters (BMEL) Vol.6 No.4
- 복사/대출신청
Purpose Image segmentation is a crucial topic in computervision and medical image processing. However, accurateimage segmentation is still a challenging task for many medicalapplications. The region growing based image segmentationprocess starts by selecting seed points within the region ofinterest. Hence, the segmentation algorithm is sensitive to theinitial seeds and the result can be influenced greatly by theaccuracy of seed selection process. Manual seed selectioncan be time-consuming and requires an expert to completethe selection. In this paper, we propose an innovativeapproach to automating the initialization process of the liversegmentation of magnetic resonance images. The seed points,which are needed to initialize the segmentation process weproposed in [1], are extracted and classified by using affineinvariant moments and artificial neural network. Methods We calculated eleven invariant moments for 56different points within the region of interest of an abdominalMR image. These points represent the bifurcation points ofthe vessels centerlines of the liver. In this paper, we dividethe shape of the liver into four regions; left hepatic vein,center hepatic vein, hepatic portal vein, and right hepaticvein. Then, the moments are classified by an artificial neuralnetwork to decide to which part of the liver each point belongs. Results We have validated our proposed technique bycomparing the method with manual seed selection. Theexperimental results show that our method outperforms themanual method in terms of the accuracy of seed pointselection and the speed of the process. Conclusions The proposed technique is considered a robusttechnique for 3D point selection and classification. Theselected seed points are used to initialize the segmentationprocess. The aim of this method is to efficiently detect andidentify the seed points in MR images.
2
Vessel-Based Fast Deformable Registration with Minimal Strain Energy

Xishi Huang,Jing Ren,Anwar Abdalbari,Mark Green 대한의용생체공학회 2016 Biomedical Engineering Letters (BMEL) Vol.6 No.1
- 복사/대출신청
Purpose Image registration for internal organs and softtissues is considered extremely challenging due to organshifts and tissue deformation caused by patients’ movementssuch as respiration and repositioning. In this paper, wepropose a fast deformable image registration method. Thepurpose of our work is to greatly improve the registrationtime while maintaining the registration accuracy. Methods In this study, we formulate the deformable imageregistration problem as a quadratic optimization problem thatminimizes strain energy subject to the constraints of 3Dcurves of blood vessel centerlines and point marks. Theproposed method does not require iteration and is localminimum free. By using 2nd order B-splines to model theblood vessels in the moving image and a new transformationmodel, our method provides a closed-form solution thatimitates the manner in which physical soft tissues deform,thus guarantees a physically consistent match. Results We have demonstrated the effectiveness of ourdeformable technique in registering MR images of the liver. Validation results show that we can achieve a targetregistration error (TRE) of 1.29 mm and an average centerlinedistance error (ACD) of 0.84 ± 0.55 mm. Conclusions This technique has the potential to significantlyimprove registration capabilities and the quality of intraoperativeimage guidance. To the best of our knowledge, thisis the first time that a global analytical solution has beendetermined for the registration energy function with 3Dcurve constraints.
3
Automatic error correction using adaptive weighting for vessel-based deformable image registration

Jing Ren,Mark Green,Xishi Huang,Anwar Abdalbari 대한의용생체공학회 2017 Biomedical Engineering Letters (BMEL) Vol.7 No.2
- 복사/대출신청
In this paper, we extend our previous work ondeformable image registration to inhomogenous tissues. Inhomogenous tissues include the tissues with embeddedtumors, which is common in clinical applications. It is avery challenging task since the registration method thatworks for homogenous tissues may not work well withinhomogenous tissues. The maximum error normallyoccurs in the regions with tumors and often exceeds theacceptable error threshold. In this paper, we propose a newerror correction method with adaptive weighting to reducethe maximum registration error. Our previous fastdeformable registration method is used in the inner loop. We have also proposed a new evaluation metric averageerror of deformation field (AEDF) to evaluate the registrationaccuracy in regions between vessels and bifurcationpoints. We have validated the proposed method using liverMR images from human subjects. AEDF results show thatthe proposed method can greatly reduce the maximumregistration errors when compared with the previousmethod with no adaptive weighting. The proposed methodhas the potential to be used in clinical applications toreduce registration errors in regions with tumors.
4
3D Surface Reconstruction of Stereo Endoscopic Images for Minimally Invasive Surgery

Jing Ren,Xishi Huang,Anwar Abdalbari 대한의용생체공학회 2013 Biomedical Engineering Letters (BMEL) Vol.3 No.3
Purpose Surface to surface registration of MR and endoscopic images is the key to MR and endoscopic image fusion,which will provide the surgeon with better 3D context of the surgical site in minimally invasive procedures. However,accurate reconstruction of 3D surface from stereo endoscopic images is still a challenging task especially for the surgical site with few features. In this paper, we propose a new method to reconstruct 3D surface from stereo endoscopic images. Methods We project a gridline light pattern onto the surgical site and then use a stereo endoscope to acquire two stereo images. The major steps in the surface reconstruction process include 1) applying an automatic method of detecting region of interest, 2) applying an image intensity correction algorithm,and 3) applying a novel automatic method to match the intersection points of the gridline pattern. Results We have validated our proposed technique on a liver phantom and compared our method with an existing method of similar scope. Our experiment results show that our method outperforms the existing method in terms of correct matching rate (98% vs. 47%) which is an indicator of the surface reconstruction accuracy. Conclusions The proposed technique has the potential to be used in clinical practice to improve image guidance in endoscope based minimally invasive procedures. This technique may also be applied to the endoscopic procedures of other organs in the abdomen, chest cavity and pelvis such as kidneys and lungs.

상세검색

RISS 보유자료

상세검색

해외전자자료

연관 검색어 추천