Difference between revisions of "Leukocyte"

From Rodinia
Jump to: navigation, search
Line 1: Line 1:
    The leukocyte application detects and tracks rolling leukocytes (white
+
The leukocyte application detects and tracks rolling leukocytes (white
 
blood cells) in in vivo video microscopy of blood vessels. The
 
blood cells) in in vivo video microscopy of blood vessels. The
 
velocity of rolling leukocytes provides important information about
 
velocity of rolling leukocytes provides important information about
Line 5: Line 5:
 
development of anti-inflammatory medications.
 
development of anti-inflammatory medications.
  
    In the application, cells are detected in the first video frame and
+
In the application, cells are detected in the first video frame and
 
then tracked through subsequent frames. Detection is accomplished by
 
then tracked through subsequent frames. Detection is accomplished by
 
computing for every pixel in the frame the maximal Gradient Inverse
 
computing for every pixel in the frame the maximal Gradient Inverse
Line 16: Line 16:
 
shape of the cell.
 
shape of the cell.
  
  Tracking is accomplished by first computing, in the area surrounding
+
Tracking is accomplished by first computing, in the area surrounding
 
each cell, a Motion Gradient Vector Flow (MGVF) matrix. The MGVF is a
 
each cell, a Motion Gradient Vector Flow (MGVF) matrix. The MGVF is a
 
gradient field biased in the direction of blood flow, and it is
 
gradient field biased in the direction of blood flow, and it is
Line 23: Line 23:
 
shape and determine the new location of each cell.
 
shape and determine the new location of each cell.
  
  For more information, see:
+
For more information, see:
 
M. Boyer, D. Tarjan, S. T. Acton, and K. Skadron. "Accelerating
 
M. Boyer, D. Tarjan, S. T. Acton, and K. Skadron. "Accelerating
 
Leukocyte Tracking using CUDA: A Case Study in Leveraging Manycore
 
Leukocyte Tracking using CUDA: A Case Study in Leveraging Manycore

Revision as of 21:20, 7 March 2009

The leukocyte application detects and tracks rolling leukocytes (white blood cells) in in vivo video microscopy of blood vessels. The velocity of rolling leukocytes provides important information about the inflammation process, which aids biomedical researchers in the development of anti-inflammatory medications.

In the application, cells are detected in the first video frame and then tracked through subsequent frames. Detection is accomplished by computing for every pixel in the frame the maximal Gradient Inverse Coefficient of Variation (GICOV) score across a range of possible ellipses. The GICOV score for an ellipse is the mean gradient magnitude along the ellipse divided by the standard deviation of the gradient magnitude. The matrix of GICOV scores is then dilated to simplify the process of finding local maxima. For each local maximum, an active contour algorithm is used to more accurately determine the shape of the cell.

Tracking is accomplished by first computing, in the area surrounding each cell, a Motion Gradient Vector Flow (MGVF) matrix. The MGVF is a gradient field biased in the direction of blood flow, and it is computed using an iterative Jacobian solution procedure. After computing the MGVF, an active contour is used once again to refine the shape and determine the new location of each cell.

For more information, see: M. Boyer, D. Tarjan, S. T. Acton, and K. Skadron. "Accelerating Leukocyte Tracking using CUDA: A Case Study in Leveraging Manycore Coprocessors." In Proceedings of the International Parallel and Distributed Processing Symposium (IPDPS), May 2009. http://www.cs.virginia.edu/~mwb7w/publications/ipdps09.pdf


Downloads