| Paper Details: | Downloads: 478 |
| Serial Number: | P1150535155
|
| Title: | Retinal Motion Detection and 3D Structure Recovery From Two Perspective Views
|
| Authors: | Tarek A. Al-Saeed and Nahed H. Solouma and Said M. El-Sherbiny
|
| Abstract: | Accurate registration, mosaicing and montage are essential for a physician employing laser photocoagulation in the treatment of diabetic choroidal neovascularization. In this paper, we introduce a 3-D based motion and structure recovery, compared to previous 2-D approaches. We detect the motion of the retina and recover its 3-D structure from vascular landmark correspondences between two perspective views. The camera was calibrated to obtain its intrinsic parameters. We propose a new technique to determine the essential matrix based on Tikhonov regularization. Two different algorithms are applied to obtain a closed-form estimate of the general 3-D rigid motion of the retina. Strict algorithms are introduced to determine the sign of the essential matrix without the need to recover the structure. Then, we verify the accuracy of our algorithm, and it gave good results. Also, we obtain the 3-D structure and then use a linear regression to fit the obtained 3-D data to the surface of a sphere.
|
| Keywords: | perspective projection, pin-hole camera, motion tracking, 3-D recovery
|
| Journal/Conference: | ICGST Conference on Graphics, Vision and Image Processing, GVIP-05
|
| Volume: |
|
| Issue: |
|
| Submission Date: | 10/19/2005 12:00:00 AM
|
| Review Date: | 11/12/2005 12:00:00 AM
|
| Publishing Date: | 12/19/2005 12:00:00 AM
|
| Article Downloads: | 478
|
| Download: |
|