Polar Run-Length Features in Segmentation of Retinal Blood Vessels

Rezatofighi, S. Hamid; Roodaki, Alireza; Pourmorteza, Amir; Soltanian‐Zadeh, Hamid

doi:10.1109/icdip.2009.18

Cited by 3 publications

(1 citation statement)

References 9 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…To reduce computational efforts, a lot of publications propose the Hessian matrix to discover the assumed vessel direction and subsequent localization with matched filters [8][9][10]. Further approaches are feature extraction and classification [13] (for example based on local binary patterns or morphological operations), line detection [14], line segment detection [6] with subsequent segment classification based on forward feature selection [15], Gabor wavelets [16,17], or thresholding [7,18]. Numerous publications apply supervised classification methods, in which, based on defined features, each pixel is classified into one of the two possible classes (blood vessel or background).…”

Section: Overview Of Related Workmentioning

confidence: 99%

Segmentation of Blood Vessel Structures in Retinal Fundus Images with Logarithmic Gabor Filters

Groß¹,

Klein²,

Schneider

2013

CMIR

View full text Add to dashboard Cite

The analysis of blood vessel structures in the retinal fundus images is important for the diagnosis of many diseases. Vessel segmentation can assist in the detection of pathological changes which are possible indicators for arteriosclerosis, retinopathy, micro aneurysms and macular degeneration. In this article, two approaches to blood vessel segmentation are presented. Both of them are based on the evaluation of phase symmetry information using complex logarithmic Gabor wavelets. In the first approach, a phase symmetry filter is combined with the front propagation algorithm fast marching, the second method uses a hysteresis thresholding step. The approaches have shown excellent results for the vessel segmentation on colon polyps. Although they were adapted to structures in retinal fundus imaging, neither eye specific knowledge nor supervised classification methods are used. For high comparability with previous publications in the field, the algorithms are evaluated on the two publicly available image databases DRIVE and STARE. The hysteresis thresholding approach which performs slightly superior achieves an average accuracy of 94.92% (sensitivity: 71.22%, specificity: 98.41%) for the DRIVE and 95, 65% (sensitivity: 71.87%, specificity: 98.34%) for the STARE database.

show abstract