| Version 4.0 (R13+) Image Processing Toolbox™ Release Notes | ![]() |
This table summarizes what's new in Version 4.0 (R13+):
| New Features and Changes | Version Compatibility Considerations | Fixed Bugs and Known Problems | Related Documentation at Web Site |
|---|---|---|---|
| Yes Details below | Yes—Details labeled as Compatibility Considerations, below. See also Summary. | Bug fixes Details | No |
New features and changes introduced in this version are
The toolbox includes a new tool for displaying images, called the Image Viewer. This tool supports zooming, scrolling, and overview navigation with large images. The Image Viewer automatically displays the pixel value at the mouse location but you can also use a special zoom tool, called the Pixel Region tool, to perform simultaneous color and quantitative inspection of individual pixels. You can also view metadata for the image file or MATLAB® variable.
To start the Image Viewer, use the imview function.
imview('board.tif')
The following figure illustrates the Image Viewer and its capabilities.
Note
On platforms that don't support Java™, have an older version
of Java, and on Macintosh® systems, calls to imview invoke the imshow function. The toolbox issues
this warning when imview is invoked: |

The toolbox includes a pair of new functions, makecform and applycform, for converting to and from a
family of standard, device-independent color spaces. The functions
support conversions between members of the family of color spaces
defined by the CIE Commission Internationale de l'Éclairage (International Commission on Illumination), including the
,
,
,
,
, and
color spaces.
The functions also support conversion to and from the industry standard
color space.
The toolbox also includes new function, iccread, for reading in ICC color profiles and using them to transform color data.
In addition, the toolbox also includes functions for converting the class representation of converted color spaces: lab2uint8, lab2uint16, lab2double, xyz2uint8, and xyz2double functions.
The toolbox includes two new image enhancement functions: adapthisteq and decorrstretch.
The adapthisteq function performs contrast-limited adaptive histogram equalization (CLAHE). This function uses a contrast-enhancement method that works significantly better than regular histogram equalization for most images.
The decorrstretch function performs a decorrelation stretch on truecolor images, or images with multiple color or spectral bands. Decorrelation stretch is a technique used to enhance, or stretch, the color differences in an image. This function can be used, for example, to aid visual interpretation when two or more bands are significantly correlated.
The dicomwrite function now supports exporting to DICOM files using the MR (magnetic resonance) and CT (computed tomography) modalities.
The dicominfo and dicomread functions can now read some files that are marginally noncompliant with the DICOM specification. Some commonly-used medical imaging devices produce such files. In addition, these functions can now read some files produced by GE® devices that use certain private transfer syntaxes.
The toolbox includes a new function, dicomuid, that generates DICOM unique identifiers. This is the same method used by the dicomwrite function.
The toolbox includes two new functions, fanbeam and ifanbeam, for computing an alternative mathematical representation of an image using fan beam projections. Using the ifanbeam function, you can reconstruct an image from fan beam projection data.
The toolbox also includes functions, fan2para and para2fan, for converting projection data between fan-beam and parallel-beam geometries. (You use the radon function to create parallel beam projection data.)
The toolbox includes a new function, bwboundaries, to trace the boundaries of all objects in a binary image. The new bwtraceboundary function traces a single boundary from a given starting point.
The toolbox includes a new function, uintlut, that changes element values in a uint8 or uint16 array by passing them through a 256-element or 65,536-element lookup table. This low-level utility function has been used to speed up other toolbox functions such as imadjust.
The image arithmetic functions have been optimized in two ways:
Portable code improvements have been made to speed up the arithmetic functions on all platforms.
Pentium- and MMX-specific code improvements have been made to provide additional speed improvements on the Windows® and Linux® platforms. The changes are based on the Intel® Performance Primitives Library.
Functions affected by these improvements include the imabsdiff, imadd, imcomplement, imdivide, imlincomb, and the immultiply functions. To determine if the Intel Performance Primitives Library is being used, call the ippl function.
A variety of existing toolbox functions have been optimized to run faster and use less memory.
Image type conversion functions
Certain image enhancement functions: imadjust and imhist
Certain color space conversion functions: rgb2gray, rgb2ntsc, rgb2ycbcr, and ycbcr2rgb
Deblurring functions: deconvblind, deconvlucy, deconvreg, and deconvwnr
In addition to the major new features, the toolbox includes several additional enhancements.
Function | Enhancement |
|---|---|
cp2tform | The algorithms have been modified to make them more robust numerically when the input coordinates have a very large offset from the origin. |
cpselect | The Control Point Selection Tool now displays a single legend window even if multiple instances of the tool are being displayed. |
imadjust | Supports a simplified, single-input syntax, in which it uses stretchlim to compute contrast-stretch parameters automatically. |
immovie | No longer flickers while a movie is being generated. |
medfilt2 | Class support has been extended to all numeric types and it uses a new algorithm for large window sizes that is significantly faster than the old one. |
ordfilt2 | Class support has been extended to all numeric types and it uses a new algorithm for large rectangular domains that is significantly faster than the old one. |
regionprops | Uses an improved method for computing the convex hull of a labeled object, resulting in more accurate ConvexHull, Convexity, and ConvexImage measurements. |
roipoly | Uses a new algorithm that produces more intuitive results and has better performance. Users who have code that requires the same output as the previous version can use the function roipolyold. |
tformfwd and tforminv | Supports additional syntaxes that make these functions easier to use for common operations. |
The sample images listed below have been removed from the toolbox.
afmsurf | bonemarr | enamel | ngc4024l | rice | testpat2 |
alumgrns | circles | flowers | ngc4024m | saturn | text |
bacteria | circlesm | ic | ngc4024s | shot1 | tissue1 |
blood1 | debye1 | lily | pearlite | testpat1 |
The following new sample images are included with the toolbox.
blobs.png | peppers.png | testpat1.png |
circles.png | rice.png | text.png |
coins.png | saturn.png | tissue.png |
glass.png | solarspectra.fts |
New ICC Profiles. The toolbox includes sample ICC profiles that can be used with the color space conversion functions.
File | Description |
|---|---|
lab8.icm | 8-bit
|
monitor.icm | Typical monitor profile |
sRGB.icm |
|
swopcmyk.icm | CMYK input profile |
The Image Processing Toolbox™, Version 4.0, includes the following bug fixes.
Function | Fix |
|---|---|
bwlabel | No longer produces a segmentation violation when called with an empty matrix. |
cpselect |
|
dicomread | Now reads DICOM files that contain 1-bit overlay data. |
imabsdiff, imadd, imdivide, imlincomb, imsubtract | The image arithmetic functions no longer error when called with logical inputs |
imfill | Now fills hole pixels on the outer edge of an image that are not connected to the background because a non-default connectivity was specified. |
imresize | The anti-aliasing filtering, applied by imresize when shrinking an image, no longer causes a narrow strip of dark pixels to appear around the edge of the image. |
label2rgb | Now uses a new default colormap. The zero-color in the previous default colormap was the same as the last color in the colormap, with the result that the object labeled with the highest number could not be distinguished from the background. |
radon | An off-by-one error in the calculation of the center pixel location for images with even dimensions has been fixed. |
rgb2ind | No longer errors when passed an input image that contain only a single color. |
strel | The syntax strel('ball', r, h, 0) returns an empty strel object instead of a degenerate ball structuring element with radius 0. |
The following functions have been obsoleted or deleted.
Function | Status |
|---|---|
isrgb | This function has been obsoleted and will issue a warning when called |
isind | This function has been obsoleted and will issue a warning when called. |
isbw | This function has been obsoleted and will issue a warning when called. |
isgray | This function has been obsoleted and will issue a warning when called |
![]() | Version 4.1 (R13SP1) Image Processing Toolbox™ | Version 3.2 (R13+) Image Processing Toolbox™ | ![]() |
| © 1984-2008- The MathWorks, Inc. - Site Help - Patents - Trademarks - Privacy Policy - Preventing Piracy - RSS |