Vessel Detection Executable

Download Command-Line Executable

This executable will process the image pixel-by-pixel to apply the Likelihood Ratio Vesselness (LRV) image wide. It is therefore slower than the vessel tracing executable.

Run

The executable has only one argument, which is the image file to process.

This executable can be used for non-commercial purposes only.

Results

Four files are written for each run: 1) Likelihood Ratio Vesselness at each pixel of the input image (header+content), 2) Vessel matched filter computed at each pixel of the input image (header+content). Each of these outputs is in the Insight Toolkit RAW format, because they are floating point outputs. Extension .mhd specifies the file header and extension .raw is the image content. Do not linearly scale the output of the LRV filter before use (i.e. by stretching to <0, 255> interval of integer intensities) because it is a likelihood ratio and you will loose important responses (many would get rounded to zero).

If you work in MATLAB, you can use the following function to read the raw images: readRaw.m
Example Usage:

readRaw( 'C:\path_to_image\imagefile.raw', [ImageWidth ImageHeight], 'float', 'native', 0 );

Note: You need to specify the image size ([ImageWidth ImageHeight]) manually. This can be determined from the header file (.mhd file).

Publications and Further Reading

When you publish results using the executable, please acknowledge the authors by citing the following paper:

1. Sofka, M., V. Stewart, C., 2006. Retinal Vessel Extraction Using Multiscale Matched Filters, Confidence and Edge Measures. IEEE Transactions on Medical Imaging 25, 1531–1546.
   abstract bibtex

   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 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 efficient and effective vessel centerline extraction algorithm.

   @article{sofka:tmi06,
     author = {Sofka, Michal and V.\ Stewart, Charles},
     title = {Retinal Vessel Extraction Using Multiscale Matched
                       Filters, Confidence and Edge Measures},
     journal = {IEEE Transactions on Medical Imaging},
     year = {2006},
     volume = {25},
     number = {12},
     pages = {1531--1546},
     month = dec,
     url = {http://www.sofka.com/vessels.html}
   }