Automatic segmentation of equine larynx for diagnosis of laryngeal hemiplegia

MD. Salehin, Lihong Zheng, Junbin Gao

Research output: Book chapter/Published conference paperConference paperpeer-review


This paper presents an automatic segmentation method for delineation of the clinically significant contours of the equine larynx from an endoscopic image. These contours are used to diagnose the most common disease of horse larynx laryngeal hemiplegia. In this study, hierarchal structured contour map is obtained by the state-of-the-art segmentation algorithm, gPb-OWT-UCM. The conic-shaped outer boundary of equine larynx is extracted based on Pascal's theorem. Lastly, Hough Transformation method is applied to detect lines related to the edges of vocal folds. The experimental results show that the proposed approach has better performance in extracting the targeted contours of equine larynx than the results of using only the gPb-OWT-UCM method.
Original languageEnglish
Title of host publicationAIP2013
Subtitle of host publicationInternational symposium on computational models for life sciences
EditorsDadong Wang Dadong Wang
Place of PublicationAustralia
PublisherAIP Publishing
Number of pages10
ISBN (Electronic)9780735411876
Publication statusPublished - 2013
EventAustralian Institute of Physics Congress - Sydney, Australia, Australia
Duration: 27 Nov 201329 Nov 2013


ConferenceAustralian Institute of Physics Congress


Dive into the research topics of 'Automatic segmentation of equine larynx for diagnosis of laryngeal hemiplegia'. Together they form a unique fingerprint.

Cite this