<html><head></head><body style="word-wrap: break-word; -webkit-nbsp-mode: space; -webkit-line-break: after-white-space; ">Hello,<div> I was noticing a problem with the boundary face calculator where some pixels were visited multiple times. </div><div>I think this could be a real problem, especially for multithreading where it is assumed that images are divided up into non-overlapping regions. I did some digging in the threads and came across two discussions, one reported the problem on October 17, 2010 (first conversation below) and another provided a fix on November 15, 2010 (second conversation below). I have taken the fix and applied it to the 3.20 branch, that patch is attached. Also, I have written a simple test driver that demonstrates the problem and attached it as well. </div><div><br></div><div>In response to the first conversation below, this is a problem for all images. Specifically, my test driver creates a 3D image that defaults to size 4x4x4, though the size can be passed in via an argument and uses a neighborhood radius of 1. I have used the test to directly confirm this is an issue for 3x3x3, 4x4x4, and 10x10x10 images.</div><div><br></div><div>Should I create a bug report for this, or simply create a gerrit patch for ITK4?</div><div><br></div><div>Thanks!</div><div>Kris</div><div><br></div><div></div></body></html>