Publication: Fuzzy zoning for feature matching technique in 3D reconstruction of nasal endoscopic images
2
Issued Date
2015-12-01
Resource Type
ISSN
18790534
00104825
00104825
Other identifier(s)
2-s2.0-84945905900
Rights
Mahidol University
Rights Holder(s)
SCOPUS
Bibliographic Citation
Computers in Biology and Medicine. Vol.67, (2015), 83-94
Suggested Citation
Surapong Rattanalappaiboon, Thongchai Bhongmakapat, Panrasee Ritthipravat Fuzzy zoning for feature matching technique in 3D reconstruction of nasal endoscopic images. Computers in Biology and Medicine. Vol.67, (2015), 83-94. doi:10.1016/j.compbiomed.2015.09.021 Retrieved from: https://repository.li.mahidol.ac.th/handle/123456789/35788
Research Projects
Organizational Units
Authors
Journal Issue
Thesis
Title
Fuzzy zoning for feature matching technique in 3D reconstruction of nasal endoscopic images
Other Contributor(s)
Abstract
© 2015 Elsevier Ltd. 3D reconstruction from nasal endoscopic images greatly supports an otolaryngologist in examining nasal passages, mucosa, polyps, sinuses, and nasopharyx. In general, structure from motion is a popular technique. It consists of four main steps; (1) camera calibration, (2) feature extraction, (3) feature matching, and (4) 3D reconstruction. Scale Invariant Feature Transform (SIFT) algorithm is normally used for both feature extraction and feature matching. However, SIFT algorithm relatively consumes computational time particularly in the feature matching process because each feature in an image of interest is compared with all features in the subsequent image in order to find the best matched pair. A fuzzy zoning approach is developed for confining feature matching area. Matching between two corresponding features from different images can be efficiently performed. With this approach, it can greatly reduce the matching time. The proposed technique is tested with endoscopic images created from phantoms and compared with the original SIFT technique in terms of the matching time and average errors of the reconstructed models. Finally, original SIFT and the proposed fuzzy-based technique are applied to 3D model reconstruction of real nasal cavity based on images taken from a rigid nasal endoscope. The results showed that the fuzzy-based approach was significantly faster than traditional SIFT technique and provided similar quality of the 3D models. It could be used for creating a nasal cavity taken by a rigid nasal endoscope.