Biomedical Background Registration

bgreg pic.jpg

Project Info

Title Background Registration for Intravital Video Microscopy of Rolling Leukocytes
Goal Design and develop an automated registration system to improve tracking rolling leukocytes in vivo
Funding National Institutes of Health (NIH)

Purpose

To develop (1) a method to compensate for "jitter" -- image translation -- caused by circulatory and respiratory systems which cause active contour trackers to lose target leukocytes and (2) a technique to register intravital video over multiple fixed and moving fields of view to allow for tracking a single cell leukocyte from the point it enters the post-capillary venule and follow it through the stage of rolling until firm adhesion to the endothelium.

Discussion

Background registration on intravital video microscopy of rolling leukocytes is needed in order to realize a completely automated and digital tracking system - a goal for Team VIVA. "Jittery" frames are common to in vivo video sequences. A jittery frame is one that is translated from the majority of frames in the sequence. These translated frames move enough between frames to interrupt tracking and cause the tracker to lose the target leukocyte.

We have developed an automated jitter compensation system using a combination of morphological techniques. The system greatly reduces jitter in the video sequences. Testing has shown promise to greatly enhance leukocyte tracking.

The following two videos available for download illustrate the correction of a jittery sequence.

Input [.avi] arrowright Output [.avi]

Below is are two examples of tracking failing due to jitter and tracking succeeding due to jitter minimization.
  • Tracker fails due to jitter [.avi]
  • Tracker Succeeds on Stabilized Sequence [.avi]

Below are two short examples of tracking over fixed and moving fields of view.

The following image shows a mosiac of the extended tracking range with the tracking path marked by red dots.

bgreg fig1.jpg
Fig 1. Mosaic of extended tracking range.
Topic revision: r3 - 30 May 2014, AndreaVaccari
 
banner.png
©2017 University of Virginia. Privacy Statement
Virginia Image and Video Analysis - School of Engineering and Applied Science - University of Virginia
P.O. Box 400743 - Charlottesville, VA - 22904 - E-Mail viva.uva@gmailREMOVEME.com