Motivated by the goals of improving detection of low-contrast and narrow vessels and eliminating false detections at non-vascular structures, a new technique is presented for extracting vessels in retinal images. The core of the technique is a new likelihood ratio test that combines matchedfilter responses, confidence measures and vessel boundary measures. Matched filter responses are derived in scale-space to extract vessels of widely varying widths. A vessel confidence measure is defined as a projection of a vector formed from a normalized pixel neighborhood onto a normalized ideal vessel profile. Vessel boundary measures and associated confidences are computed at potential vessel boundaries. Combined, these responses form a 6-dimensional measurement vector at each pixel. A training technique is used to develop a mapping of this vector to a likelihood ratio that measures the "vesselness" at each pixel. Results comparing this vesselness measure to matched filters alone and to measures based on the Hessian of intensities show substantial improvements both qualitatively and quantitatively. The Hessian can be used in place of the matched filter to obtain similar but less-substantial improvements or to steer the matched filter by preselecting kernel orientations. Finally, the new vesselness likelihood ratio is embedded into a vessel tracing framework, resulting in an e#cient and e#ective vessel centerline extraction algorithm.

Motivated by the goals of improving detection of low-contrast and narrow vessels and eliminating false detections at non-vascular structures, a new technique is presented for extracting vessels in retinal images. The core of the technique is a new likelihood ratio test that combines matched filter responses, confidence measures and vessel boundary measures. Matched filter responses are derived in scale-space to extract vessels of widely varying widths. A vessel confidence measure is defined as a projection of a vector formed from a normalized pixel neighborhood onto a normalized ideal vessel profile. Vessel boundary measures and associated confidences are computed at potential vessel boundaries. Combined, these responses form a 6-dimensional measurement vector at each pixel. A learning technique is applied to map this vector to a likelihood ratio that measures the "vesselness" at each pixel. Results comparing this vesselness measure to matched filters alone and to measures based on the intensity Hessian show substantial improvements both qualitatively and quantitatively. When the Hessian is used in place of the matched filter, similar but less-substantial improvements are obtained. Finally, the new vesselness likelihood ratio is embedded into a vessel tracing framework, resulting in an e#cient and e#ective vessel extraction algorithm.

We propose an adaptive detection scheme for large and small blood vessels in color retinal images. Our scheme consists of three functions: adaptive contrast enhancement, feature extraction of blood vessels, and tracing. The average performance of our method over twenty tested images is 84.3 % for true positive rate (TPR), and 3.9 % for false positive rate (FPR). For normal images, the TPRs range from 80% to 91%, and their corresponding FPRs range from 2.8 % to 5.5%. For abnormal images, the TPRs range from 73.8 % to 86.5 % and the FPRs range from 2.1 % to 5.3%, respectively. Small vessels take up 42 % of overall vessel pixels, where 75 % of small vessels were captured by our method. Enhancement of blood vessels is achieved from the extension of the adaptive histogram equalization technique. Feature extraction of small blood vessels is by using the standard deviation of Gabor filter responses along different orientations. Tracing of the vascular network consists of three major functions: forward detection, backward verification, and bifurcation detection. Combining extrapolation and local greedy search reduces the prediction errors of vessel directions by 15-20%. Only two sample images and their hand-traced maps are needed for parameter training and calibration. Index terms--- Retinal images, adaptive contrast enhancement, blood vessel tracing, Gabor filter.