</ul><a name="match"></a><hr><h4>Matching</h4><ul><li type=square><ahref=Match/matchbycorrelation.m>matchbycorrelation.m</a> generatesputative matches between previously detected feature points in twoimages by looking for points that are maximally correlated with eachother within windows surrounding each point.  Only points thatcorrelate most strongly with each other in both directions arereturned.  This is a simple-minded N<sup>2</sup> comparison.<li type=square><ahref=Match/matchbymonogenicphase.m>matchbymonogenicphase.m</a> issimilar to matchbycorrelation, but instead matches on oriented phasevalues rather than greyscale values.  This matcher performs ratherwell relative to normalised greyscale correlation.  Typically thereare more putative matches found and fewer outliers.  There is agreater computational cost in the pre-filtering stage but potentiallythe matching stage is much faster as each pixel is effectively encodedwith only 3 bits. (Though this potential speed is not realized in thisimplementation). See <a href=Robust/example/testfund.m>testfund</a>below to see an example of the use of this function.</ul><a name="robust"></a><hr><h4>Model fitting and Robust estimation</h4><center><table><tr><td><img align=top src=Robust/example/putativesm.jpg><br>Putative matches obtained<br>by <a href=Match/matchbycorrelation.m>matchbycorrelation.m</a></td><td>&nbsp; &nbsp;</td><td><img align=top src=Robust/example/inlierssm.jpg><br>Inlying matches consistent<br>with fundamental matrix</td></tr></table></center><ul><li type=square><a href=Robust/ransac.m>ransac.m</a> a general purposeimplementation of the RANSAC algorithm.  <imgsrc="http://www.cs.uwa.edu.au/images/updated.gif"> August 2008 Thisshould be fixed now - sorry for any problems you had<br><li type=square><a href=Robust/ransacfithomography.m>ransacfithomography.m</a>robustly fits a homography to a set of putatively matched image points. <li type=square><ahref=Robust/ransacfitfundmatrix.m>ransacfitfundmatrix.m</a> robustlyfits a fundamental matrix to a set of putatively matched image points.This function uses an 8 point fundamental matrix solution.<li><ahref=Robust/ransacfitfundmatrix7.m>ransacfitfundmatrix7.m</a> robustlyfits a fundamental matrix to a set of putatively matched image points.This function requires Andrew Zisserman's 7 point fundamental matrix codeavailable from: <a href="http://www.robots.ox.ac.uk/~vgg/hzbook/code/">MATLAB Functions for Multiple View Geometry</a><li type=square><ahref=Robust/ransacfitaffinefund.m>ransacfitaffinefund.m</a> robustlyfits an affine fundamental matrix to a set of putatively matched imagepoints.<li type=square><a href=Robust/ransacfitplane.m>ransacfitplane.m</a>robustly fits a plane to 3D data points.<li type=square><a href=Robust/ransacfitline.m>ransacfitline.m</a>robustly fits a line to 3D data points.<li type=square><a href=Robust/iscolinear.m>iscolinear.m</a> tests if3 points are colinear, used by ransacfitplane and ransacfithomography.<li type=square><a href=Robust/fitline.m>fitline.m</a> least squaresfit of a line to 2D data points.<li type=square><a href=Robust/fitline3d.m>fitline3d.m</a> least squaresfit of a line to 3D data points. Contributed by Felix Duvallet.<li type=square><a href=Robust/fitplane.m>fitplane.m</a> least squaresfit of a plane to 3D data points.<li type=square><a href=Robust/example/testfitplane.m>testfitplane</a>example of using ransacfitplane.m<li type=square><a href=Robust/example/testfitline.m>testfitline</a>example of using ransacfitline.m<li type=square><a href=Robust/example/testfund.m>testfund</a>example of using ransacfitfundmatrix.m<li type=square><a href=Robust/example/testhomog.m>testhomog</a>example of using ransacfithoography.m<li type=square><a href=Robust/randomsample.m>randomsample</a> a basicreplacement for <tt>randsample</tt> to be used with <tt>ransac.m</tt>should you not have the statistics toolbox, or are using Octave.<li><a href=Robust/example/index.html>Example</a> of using thesefunctions above to find the fundamental matrix.  </ul><!--<a name="lensdistort"></a><hr><h4>Lens Distortion Correction</h4><center><table><tr><td><img align=top src=WWWImages/distortsm.jpg><br><center>original image</center></td><td><img align=top src=WWWImages/correctedsm.jpg><br><center>corrected image</center></td></tr></table></center><p>Sorry, this code is broken due to changes I have made to the edgelinking code above.  Hopefully I will get round to fixing it soon.<ul><li><a href=LensDistort/lensparam.m>lensparam.m</a> Estimates lensdistortion parameters.  This is an implementation of Devernay andFaugeras' technique for lens distortion removal from structuredscenes.  Edges are detected in the image and line segments fitted tothe edges.  If an edge is curved it will deviate from the fittedsegment.  An optimisation process is performed on the lens distortionparameters in order to minimize the deviation of the edge points fromstraight lines.<li><a href=LensDistort/imlenstransform.m>imlenstransform.m</a> Thisfunction applies or inverts a lens distortion transformation definedby first order coefficients to an image.<li><a href=LensDistort/crookedness.m>crookedness.m</a> Returns ameasure of the 'crookedness' of a set of lines. This is the objectivefunction passed to the optimisation function by lensparam.m.<li><ahref=LensDistort/edgelistlenscorrect.m>edgelistlenscorrect.m</a> Thisfunction corrects for lens distortion defined by first ordercoefficients on an edgelist.<li><a href=LensDistort/example/index.html>Example</a> of the use ofthese functions.  </ul><p>These functions make use of:<br>drawedgelist.m, edgelink.m, lineseg.m, maxlinedev.m, mergeseg.m, implace.m, removenan.m, show.m, fitline.m--><a name="fingerprints"></a><hr><h4>Fingerprint enhancement</h4><center><table><tr><td><img align=top src=WWWImages/fingervsm.png></td><td>&nbsp; &nbsp;</td><td><img align=top src=WWWImages/finalvsm.png></td></tr></table></center><ul><li><a href=FingerPrints/ridgesegment.m>ridgesegment.m</a> identifiesridge-like regions of a fingerprint image.  It also normalises theintensity values of the image.<li><a href=FingerPrints/ridgeorient.m>ridgeorient.m</a> estimates thelocal orientation of ridges in a fingerprint.<li><a href=FingerPrints/plotridgeorient.m>plotridgeorient.m</a> plotsridge orientations calculated by ridgeorient.<li><a href=FingerPrints/ridgefreq.m>ridgefreq.m</a> estimates thelocal ridge frequency across a fingerprint image.<li><a href=FingerPrints/freqest.m>freqest.m</a> estimates the ridgefrequency within a small block of an image.  This is used byridgefreq.<li><a href=FingerPrints/ridgefilter.m>ridgefilter.m</a> enhancesa fingerprint image using oriented filters.<li><a href=FingerPrints/Docs/index.html>Example</a> of using these functions above.</ul><hr> <h4>Interesting images </h4><center><img align=top src=WWWImages/clouds.gif></center><br> <ul><li type=square><a href=Misc/noiseonf.m>noiseonf.m</a> generates noiseimages with specified amplitude spectra.  One can create pleasingcloud pattern images this way.<li type=circle> <a href=Misc/cloud9.m>cloud9.m</a> creates a moviesequence of noise images with specified amplitude spectra.  Veryrelaxing.</ul>    <br><br><center><img align=top src=WWWImages/derespk.gif></center> <br><ul><li type=square><a href=Misc/derespolar.m>derespolar.m</a> Generatesderesolved images in polar coordinates.<li><a href=Misc/polartrans.m>polartrans.m</a> Generates a polartransformation of an image.  A linear or logarithmic radiustransformation can be specified.<li type=square><a href=FrequencyFilt/circsine.m>circsine.m</a>Generates a test image consisting of a circular sine wave grating.</ul>     <hr><h4>ASCII Image Generation</h4><center> <img align=top src=WWWImages/MonaASCIIsml.gif></center> <br><ul><li type=square><a href=Misc/matscii.m>matscii.m</a> Function togenerate ASCII images from grey scale images. A bit retro but may haveapplications for interesting text messages on your mobile phone!</ul>     <hr><a name="display"></a><h4>Image Display, Image Writing and Miscellaneous</h4><ul><li type=square><a href=Misc/show.m>show.m</a> This function displaysan image, automatically setting the colour map to grey if it is a 2Dimage, or leaving it as colour otherwise, and setting the axes to be'equal'.  The image is also displayed as 'TrueSize', that is, pixelson the screen match pixels in the image.  Finally a titlecorresponding to the image variable name is added.<li type=square><a href=Misc/showfft.m>showfft.m</a> displays theamplitude spectrum of an fft.<li type=square><a href=Misc/showlogfft.m>showlogfft.m</a> displaysthe log amplitude spectrum of an fft.<li type=square><a href=Misc/imwritesc.m>imwritesc.m</a> This functioncombines image rescaling and writing into the one function.  If theimage type is <tt>double</tt> image values are rescaled to the range0-1 so that no overflow occurs when writing 8-bit intensity values.The image format to use is determined by MATLAB from the file ending.If the image type is of <tt>uint8</tt> no rescaling is performed.<li type=square><a href=Misc/matprint.m>matprint.m</a> This functionprints out a matrix using a specified C style format string.  Oftenyou find that MATLAB's default number formats are not what you want.It is surprising that there is no convenient fine control of outputformat.  I hope this function helps.<li type=square><a href=Misc/removenan.m>removenan</a> Replaces NaNvalues in a matrix with a specified default value.  Useful when youwant to prevent NaNs from contaminating and destroying some operationon an array, for example, an FFT.<li><a href=Misc/fillnan.m>fillnan</a> Replaces NaNvalues in a matrix with the value in the closest non-Nan pixel.<li type=square><a href=Misc/digipts.m>digipts.m</a> Function todigitise points in an image.  This function uses the cross-hair cursorprovided by GINPUT.  I find this is much more useable than the cursorused by IMPIXEL. In addition each location digitised is marked with ared '+'.<li type=square><a href=Misc/implace.m>implace.m</a> Function to placean image at a specified location within a larger image.</ul><hr><!-- Start of StatCounter Code --><script type="text/javascript" language="javascript">var sc_project=2025066; var sc_invisible=1; var sc_partition=18; var sc_security="bc8bdd58"; </script><script type="text/javascript" language="javascript" src="http://www.statcounter.com/counter/counter.js"></script><noscript><a href="http://www.statcounter.com/" target="_blank"><img  src="http://c19.statcounter.com/counter.php?sc_project=2025066&java=0&security=bc8bdd58&invisible=1" alt="web hit counter" border="0"></a> </noscript><!-- End of StatCounter Code --></BODY></HTML>