<p>In <a href="http://homepages.inf.ed.ac.uk/rbf/CVonline/LOCAL_COPIES/TUZEL1/MeanShift.pdf">http://homepages.inf.ed.ac.uk/rbf/CVonline/LOCAL_COPIES/TUZEL1/MeanShift.pdf</a>,<br>
It seems kernel must be radially symmetric.<br>
(using the notation in the paper above) What I use to calculate mean shift vector is the derivative of  k(s) rather than the derivative of K(x). <br>
For Gaussian kernel (u=1), dk(s)/ds = -0.5 * exp(-0.5 * s), dK(x)/dx = -x * exp(-0.5 * x). <br>
But for the function Gradient() in Gaussian Kernel class, I think it should implement dK(x)/dx.<br>
It may be strange for Gradient() to implement dk(s)/ds.<br>
I think it should have another name, but I really don't know what name is appropriate.</p>

<p style="font-size:small;-webkit-text-size-adjust:none;color:#666;">&mdash;<br>Reply to this email directly or <a href="https://github.com/mlpack/mlpack/pull/388#issuecomment-72311878">view it on GitHub</a>.<img alt="" height="1" src="https://github.com/notifications/beacon/AJ4bFAXtPJmGABQBy4gxzXvY3_uhfmaRks5nnJ_1gaJpZM4DTzb1.gif" width="1" /></p>
<div itemscope itemtype="http://schema.org/EmailMessage">
  <div itemprop="action" itemscope itemtype="http://schema.org/ViewAction">
    <link itemprop="url" href="https://github.com/mlpack/mlpack/pull/388#issuecomment-72311878"></link>
    <meta itemprop="name" content="View Pull Request"></meta>
  </div>
  <meta itemprop="description" content="View this Pull Request on GitHub"></meta>
</div>