advertisement
Abstract #60441 Published in IGR 16-4
Reliable feature selection for automated angle closure glaucoma mechanism detection
Niwas SI; Lin W; Bai X; Kwoh CK; Sng CC; Aquino MC; Chew PTJournal of Medical Systems 2015; 39: 21
Glaucoma is an eye disease where a loss of vision occurs as a result of progressive optic nerve damage usually associates with high intraocular pressure. A subtype of glaucoma called primary angle-closure glaucoma (PACG) has been observed to be the result of one or more mechanisms such as Pupil block, Plateau iris, Peripheral iris roll, and Lens in the anterior segment of the eye. Reliable features in anterior segment images are important for determining the specific mechanisms involved in PACG. In this paper, first the discriminant features are selected by several feature selection algorithms in the context of PACG detection based on anterior segment optical coherence tomography (AS-OCT) images, and then a novel criteria is proposed to further select more reliable features. Our approach is based on selecting the top-ranked features in each algorithm and its rank combination for selection of the best features. Compared with the features selected by the individual feature selection methods, the features selected by our method achieves the best performance in terms of the accuracy of classification of the four PACG mechanisms by using AdaBoost classifier.
School of Computer Engineering, Nanyang Technological University (NTU), Singapore, 639798, Singapore, issacniwas@ntu.edu.sg.
Full articleClassification:
9.3.5 Primary angle closure (Part of: 9 Clinical forms of glaucomas > 9.3 Primary angle closure glaucomas)
6.9.2.1 Anterior (Part of: 6 Clinical examination methods > 6.9 Computerized image analysis > 6.9.2 Optical coherence tomography)
