Mercurial > forge
view main/image/NEWS @ 11279:442cb00e506b octave-forge
rgbplot: add new style option to also display composite colormap
author | carandraug |
---|---|
date | Sun, 02 Dec 2012 16:57:32 +0000 |
parents | 98dbd58ba857 |
children | a09d86cf7251 |
line wrap: on
line source
Summary of important user-visible changes for image 2.0.1: ------------------------------------------------------------------- ** The following functions are new: checkerboard ** The plot produced by `imhist' is correctly scaled on the X axis so that the colorbar corresponds to the actual intensity of the stems; the given colormarp is used on the colorbar for indexed images; and the stems no longer display the markers at their top. The Y axis is also adjusted in case of peaks with high values that prevent a good overview of the histogram. ** The `fftconv2' usage "fftconv2 (v1, v2, a)", for convolution of a matrix by one vector in the column direction and another vector in the row direction, has been removed. Please use conv2 for such cases. ** The function `rgbplot' accepts a style option which alows for a composite display of a colormap. Summary of important user-visible changes for image 2.0.0: ------------------------------------------------------------------- ** The following functions are new: analyze75info imabsdiff iptcheckconn analyze75read imadd iptcheckmap analyze75write imbothat iptchecknargin blockproc imcrop iptcheckstrs bwlabeln imdivide iptnum2ordinal getrangefromclass imlincomb iscolormap im2int16 immultiply normxcorr2 im2single imsubtract wavelength2rgb ** The following functions have been deprecated in previous releases of the image package and have now been removed: imginfo ** The function `deriche' has been removed. ** The complete set of functions to work with Analyze 7.5 files has been implemented. See `analyze75info', `analyze75read' and `analyze75write'. ** `graythresh' can optionally accept an histogram rather than an image. This allows for preprocessing of the histogram previous to an automatic threshold selection. ** Otsu's method for automatic threshold selection (default for `graythresh') has been completely rewritten and should perform faster. Now, it can also return a second value representing the ``goodness'' of the computed threshold value (within class variance). ** Alternative algorithms for automatic threshold have been implemented in `graythresh' (thanks to Antti Niemistö for releasing HistThresh toolbox http://www.cs.tut.fi/~ant/histthresh/ from where many were ported, under a GPL license). Currently, the following algorithms have been implemented (see graythresh for notes and references): concavity MaxEntropy minimum Otsu intermeans MaxLikelihood MinError percentile intermodes mean moments ** The following functions have been deprecated (see their help text for the recommended alternatives): blkproc bmpwrite dilate erode ** With the new function `imbothat' the transform option of `imtophat' has been deprecated. ** The following functions have had been changed for bug fixes and/or improved matlab compatibility bwarea imhist im2uint8 isind bweuler imnoise im2uint16 mat2gray bwfill conndef isbw rgb2gray cmpermute im2bw isgray cmunique im2double isrgb ** `bwarea' now supports all image classes and considers objects all non zero pixels (not all pixels higher than zero). ** `rgb2gray' now also supports images of the class single and performs a weighted conversion to keep the image luminance instead of a the mean through each color. ** `im2bw' now supports input images of the int16 class and deals better with RGB images since it uses `rgb2gray' internally (see changes to rgb2gray). Threshold is performed on all values greater than value instead of greater than or equal. ** `imhist' is much more compatible with matlab and among other changes, it now uses the whole range of the class for the histogram rather than the minimum and maximum of the input image and displays a colorbar under the histogram. ** `isbw' now defines a black-and-white image as a binary non-sparse matrix. This is compatible with matlab. To use the old behaviour, use the new option for the call "isbw (img, "non-logical"). For backwards compatibility, if a non-logical matrix of 0 and 1 is used as input, `isbw' will still return true but a warning will be issued since this will deprecated later. ** `isgray' now also returns true for matrices of the int16 class. ** `isrgb' now returns false for logical matrix. ** `tiff_tag_read' had several bug fixes and can now check IFDs beyond the first. It can also accept mutiple tag values and IFDs simultaneously and return a matrix of the values found. Its documentation has been expanded (as well as an explanation of TIFF structure on the source) ** For sake of matlab compatibility, the behaviour of `mat2gray' has been greatly changed. Among the changes, it will no longer swap the minimum and maximum options if the first is larger than the later. Instead, will return the image complement after truncation. Also, when the maximum and minimum values are equal, `mat2gray' will truncate all values between 0 and 1. See the help text (or source) for a detailed description of cautions. ** `bwfill' was fixed to always returns a logical matrix. ** `imnoise' has been expanded to accept images of differente classes instead of only double and single. ** The private function `__bwdist` has been renamed `__bwdist__` ** Package is now dependent on GNU Octave version 3.6.0 or later. ** Package is now dependent on the signal package version 1.2.0 or later. ** Package is no longer automatically loaded.