Natarajan Sudha

Nanyang Technological University, Tumasik, Singapore

Are you Natarajan Sudha?

Claim your profile

Publications (2)0 Total impact

  • Fahad Hameed Ahmad · N. Sudha ·
    [Show abstract] [Hide abstract]
    ABSTRACT: Image Guided Interventions (IGI) has the potential to replace conventional open and invasive procedures. Such interventions also minimize damage to healthy tissue. Real-time 3D visualization of the problem area is now becoming a necessity for successful IGI. Real-time intra-operative imaging devices like multi-slice CT or 3D ultrasound are noisy and have low resolution, hence they must be compared with pre-operative high resolution images like MRI. Registering and comparing multi-modal images is inherently difficult due to the differences in visualization and the large number of missing features. A new deformable rigid body technique is presented that uses differential geometry based features and the Hausdorff similarity measure for rigid body sub-division registration. The differential geometry based features and the Hausdorff similarity measure are chosen because of their robustness to missing features and intensity variations. A translation deformation field is obtained that maps each pixel to a new location, followed by a cubic interpolator to obtain the registered image. Experimental results show that the new registration technique offers better accuracy and quality and is computationally less intensive compared with previous methods.
    01/2010; 19(4):363-377. DOI:10.1515/JISYS.2010.19.4.363
  • Fahad Hameed Ahmad · Natarajan Sudha · Jimmy Jiang Liu ·
    [Show abstract] [Hide abstract]
    ABSTRACT: A new weighted quaternion based non-rigid registration is presented in this paper. Strong crest points derived from principal curvatures provide the most robust features for image registration. Crest point strengths are based on their principal curvatures and the number of scales a particular crest point is detected at. Geometric features are extracted which are invariant to rotation, translation and scaling by using neighborhood crest points only as other voxels are susceptible to deformation. The neighborhood size is adjusted according to scale adaptively using a fixed k nearest neighbor to make the extracted feature scale invariant. Statistical properties are used to measure the distribution of these geometric invariant features. The scale and rotation invariant feature points are then used to establish a point to point correspondence between the template crest points and the subject image crest points. A multi-scale feature based subdivision scheme is employed for registration where a weighted quaternion matrix provides a quaternion transformation based on the corresponding points to obtain the best rotation for global as well as local sub-blocks.
    Medical Biometrics, Second International Conference, ICMB 2010, Hong Kong, China, June 28-30, 2010. Proceedings; 01/2010

Top Journals


  • 2010
    • Nanyang Technological University
      • School of Computer Engineering
      Tumasik, Singapore