Changes to DIPimage
- We have made some changes for compatibility with the new graphics system of MATLAB R2014b. Previous versions of DIPimage cannot be used with versions of MATLAB R2014b and newer, but this version of DIPimage does also work with older versions of MATLAB.
- New functions:
- There is a new option in the 'File' menu of the DIPimage GUI to record macros. All commands executed through the GUI will be saved to an M-file (MATLAB script). The macro can be executed simply by typing the file's name at the command prompt.
- countingframe implements a counting frame: it determines which objects in the image to count (or measure) to avoid a bias due to objects partially in the image.
- xx1, yy1, zz1 and ramp1 are one-dimensional versions of the corresponding functions xx, yy, zz and ramp. With the new broadcasting operations in DIPimage (singleton expansion), these can save a lot of memory. For example, xx1(im)^2+yy1(im)^2 produces the same result as xx(im)^2+yy(im)^2, but is much faster and uses less memory.
- Changed functions:
- When saving a displayed image through the 'Save As...' menu option of the image display window, the default file format now is PNG.
- gaussianblob has been rewritten to allow multiple blobs also in dimensionalities other than 2. The sigma parameter now can specify one sigma for each dimension and one sigma for each blob (this used to be: one sigma for each dimension if one blob was drawn, one sigma per blob if multiple blobs were drawn).
- clip now works with tensor and color images.
- The function colorspace has been rewritten. It is now much easier to add new color spaces.
- Bug fixes:
- dipshow(N,img) would create a new window if no window with handle N existed, but the new window's handle would not necessarily be N.
- The lut function once again uses the data type of the look-up table as the data type of the output image. It now also preserves singleton dimensions.
- readtimeseries now works well again when the files being read contain color images.
- The 'background' mode of threshold would fail in some situations.
- When converting from a numeric array to a dip_measurement object, the 'axes' and 'units' elements were wrong.
- Fixed documentation for functions inner and outer.
Changes to DIPlib
(These might or might not propagate to DIPimage)
- Bug fixes:
- dip_Clip() did not work correctly for images with data type dip_sint32.
- Some fixes to the documentation of dip_FindShift().
Changes to dipIO
(This propagates to DIPimage)
- Bug fixes:
- dipio_ImageReadColourSeries() did strange things when the files being read contained color images.