Feature Extraction for Classification from Images: A Look at the Retina

Author

Chui, Foon Chi Francis ; Bindoff, Ivan ; Williams, Raymond ; Byeong Ho Kang

Author_Institution

Univ. of Tasmania, Hobart, TAS

fYear

2008

fDate

13-15 Oct. 2008

Firstpage

93

Lastpage

98

Abstract

A method is proposed for the representation of localised features using disjoint sub-images taken from several datasets of retinal images for use within an incremental learning system. A tile-based localised adaptive threshold selection method was taken for vessel segmentation based on separate colour components. Arteriole-venous differentiation was done using the composite of these components and high quality fundal images. Vessel segmentation performance was evaluated on the DRIVE and STARE datasets achieving average specificity of 0.9379 and sensitivity of 0.5924.

Keywords

eye; feature extraction; image classification; image colour analysis; image representation; image segmentation; learning (artificial intelligence); medical image processing; DRIVE dataset; STARE dataset; arteriole-venous differentiation; colour component; feature extraction; image classification; image representation; incremental learning system; quality fundal image; retinal image; tile-based localised adaptive threshold selection method; vessel segmentation; Biomedical imaging; Biomedical optical imaging; Blood; Clinical diagnosis; Feature extraction; Image segmentation; Learning systems; Optical sensors; Retina; Veins; analysis; drive; fundal; image; mcrdr; retina; retinal; segmentation; stare; vessel;

fLanguage

English

Publisher

ieee

Conference_Titel

Ubiquitous Multimedia Computing, 2008. UMC '08. International Symposium on

Conference_Location

Hobart, ACT

Print_ISBN

978-0-7695-3427-5

Type

conf

DOI

10.1109/UMC.2008.27

Filename

4656524