<div dir="ltr">yes of course, I did not paid attention. sorry :-D<div><br></div><div><div><br></div><div>I have been using Fitzgibbon, Pilu and Fischer solution for fitting ellipses, which is states the problem such that it can be written as and eigenvalue equation (it is not iterative). Please, take a look for example at</div></div><div><br></div><div><a href="http://nicky.vanforeest.com/misc/fitEllipse/fitEllipse.html">http://nicky.vanforeest.com/misc/fitEllipse/fitEllipse.html</a><br></div><div><br></div><div>The original paper I think it is:</div><div><a href="http://cseweb.ucsd.edu/~mdailey/Face-Coord/ellipse-specific-fitting.pdf">http://cseweb.ucsd.edu/~mdailey/Face-Coord/ellipse-specific-fitting.pdf</a><br></div><div><br></div><div>regards.</div><div><br></div><div>F</div></div><div class="gmail_extra"><br clear="all"><div><div class="gmail_signature" data-smartmail="gmail_signature"><div>..-. ..-.<br>Fabricio Ferrari [<a href="http://www.ferrari.pro.br/" target="_blank">www.ferrari.pro.br</a>]<br>IMEF - FURG<br>Rio Grande, RS<br>Brasil</div></div></div>
<br><div class="gmail_quote">2017-02-15 10:47 GMT-02:00 Andrzej Szary <span dir="ltr"><<a href="mailto:aszary@gmail.com" target="_blank">aszary@gmail.com</a>></span>:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div dir="ltr"><div>Dear Fabricio,</div><div><br></div><div>Note that the DATA and MODEL parameters are different. For a good fit they should be similar. You can get better results using different fitter (<b>fi = fitting.SimplexLSQFitter()</b>), but it is still not good enough for my purpose (I need to find theta parameter). </div><div><br></div><div>Best wishes<br>Andrzej Szary</div><div><br></div><div><br></div><div>>> On 15 February 2017 at 13:14, Fabricio Ferrari <<a href="mailto:fabricio.ferrari@gmail.com" target="_blank">fabricio.ferrari@gmail.com</a>> wrote:</div><div>>> </div><div>>> Dear Andrzej</div><div>>> </div><div>>> here it seemed to work. Maybe the version you are using.</div><div>>></div><div>>> Here I have:</div><div>>><br></div><div>>>In [2]: np.__version__</div><div>>>Out[2]: '1.11.1'</div><div>>><br></div><div>>>In [4]: astropy.__version__</div><div>>>Out[4]: u'1.2.1'</div><div>>></div><div>>>My output:</div><div>>><br></div><div>>>DATA:</div><div>>>Model: Ellipse2D</div><div>>>Inputs: (u'x', u'y')</div><div>>>Outputs: (u'z',)</div><div>>>Model set size: 1</div><div>>>Parameters:</div><div>>> amplitude x_0 y_0 a b theta</div><div>>> --------- --- --- --- --- -----</div><div>>> 1.0 0.0 0.0 2.0 1.0 0.0 </div><div>>><br></div><div>>><br></div><div>>>MODEL:</div><div>>>Model: Ellipse2D</div><div>>>Inputs: (u'x', u'y')</div><div>>>Outputs: (u'z',)</div><div>>>Model set size: 1</div><div>>>Parameters:</div><div>>> amplitude x_0 y_0 a b theta</div><div>>> -------------- --- --- --- --- -----</div><div>>> 0.338211382114 0.1 0.1 3.0 2.0 0.2 </div><div>>><br></div><div>>>p.s. thank you for letting me know about the Ellipse2D fit, that I indeed need and did not >>know about.</div><div>>><br></div><div>>>Regards</div><div>>>Fabricio<br></div><div><br></div></div><div class="gmail_extra"><br><div class="gmail_quote">On 8 February 2017 at 16:40, Andrzej Szary <span dir="ltr"><<a href="mailto:aszary@gmail.com" target="_blank">aszary@gmail.com</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div class="m_-8964541455535655336HOEnZb"><div class="m_-8964541455535655336h5"><div dir="ltr"><div>Hello,</div><div><br></div><div>I am struggling with fitting an ellipse using Ellipse2D model. The fitting procedure does not seem to have any effect on ellipse parameters (maybe except amplitude). </div><div><br></div><div>Please find a test code at the end of this email.</div><div><br></div><div>Best wishes,</div><div>Andrzej Szary</div><div><br></div><div><br></div><div>import numpy as np</div><div>from astropy.modeling import models, fitting</div><div>import matplotlib.pyplot as pl</div><div><br></div><div># data</div><div>num = 100</div><div>x, y = np.meshgrid(np.linspace(-5., 5., num), np.linspace(-5, 5, num))</div><div>e0 = models.Ellipse2D(amplitude=1., x_0=0., y_0=0., a=2, b=1, theta=0.)</div><div>z0 = e0(x, y)</div><div>print 'DATA:\n', e0, '\n\n'</div><div><br></div><div># initial model</div><div>ei = models.Ellipse2D(amplitude=1., x_0=0.1, y_0=0.1, a=3, b=2, theta=0.2)</div><div>fi = fitting.LevMarLSQFitter()</div><div><br></div><div># fitted model?</div><div>e1 = fi(ei, x, y, z0)</div><div>z1 = e1(x, y)</div><div>print 'MODEL:\n', e1, '\n\n'</div><div><br></div><div>pl.imshow(z0, extent=[-5, 5, -5, 5], alpha=0.5)</div><div>pl.imshow(z1, extent=[-5, 5, -5, 5], alpha=0.2)</div><div>pl.show()</div><div><br></div></div>
</div></div></blockquote></div><br></div>
</blockquote></div><br></div>