Version 3.2 (R13+) Image Processing Toolbox™

This table summarizes what's new in Version 3.2 (R13+):

New Features and ChangesVersion Compatibility ConsiderationsFixed Bugs and Known ProblemsRelated Documentation at Web Site
Yes
Details below
Yes—Details labeled as Compatibility Considerations, below. See also Summary.NoNo

New features and changes introduced in this version are

Writing DICOM Files

The Image Processing Toolbox™ now supports writing files in Digital Imaging and Communications in Medicine (DICOM) format, using the dicomwrite function. Previous releases of the toolbox supported reading DICOM files with the dicomread function and reading metadata from a DICOM file using the dicominfo function.

Representing Binary Images

In previous releases, toolbox functions that returned binary images returned them as uint8 logical arrays. The toolbox used the presence of the logical flag to signify that the data range in the file was [0,1].

With this release, the toolbox returns binary images as logical arrays, using the new MATLAB® logical data type. For more information about the new logical class, see the MATLAB 6.5 Release Notes.

Compatibility Considerations

The change in the representation of binary images has the following compatibility considerations.

Change to Data Type of Output Binary Images.   All the Image Processing Toolbox functions that return a binary image now return a binary image of class logical. In previous releases, these functions returned binary images of a numeric class with the logical flag set. The Image Processing Toolbox used the existence of the logical flag to identify a binary image.

If your application checks the data type of the binary images returned by Image Processing Toolbox functions, you will need to change your code.

Change to Interpretation of Input Images.   Image Processing Toolbox functions that accept different types of images, such as grayscale and binary, no longer attempt to determine if an input image of a numeric class is intended to be a binary image.

In previous releases, toolbox functions that accepted different types of images checked the contents of an image to determine how to interpret it. For example, if an image was of class double and contained only 0s and 1s, the toolbox function would interpret it as a binary image. With Version 3.2, the toolbox only interprets images of class logical as binary images.

In the Image Processing Toolbox, the names of functions that accept both grayscale and binary images typically start with the characters "im", such as imdilate.

Converting Binary Images to an Integer Data Type.   With this release, if you convert a binary image to a numeric type, the image ceases to be a binary image.

In previous releases, the Image Processing Toolbox conversion functions im2uint8 and im2double preserved the binary attribute of the converted image. For example, if you converted a binary image of class double, which had the logical flag set, the output image returned by the im2uint8 function would also be a logical image of class uint8, with the logical flag set.

For example, create a simple logical array

bw = logical([1 0; 0 1])
bw =

     1     0
     0     1
whos
Name           Size                   Bytes  Class

bw             2x2                        4  logical array

When you convert this array to a uint8 data type, notice that it is no longer of class logical.

new_image = im2uint8(bw)

new_image =

   255     0
     0   255

whos
Name           Size                   Bytes  Class

bw               2x2                      4  logical array
new_image        2x2                      4  uint8 array

Better Input Image Type Checking

The toolbox now performs more error checking of input images, specifically input classes, attributes and option string processing, with clearer error messages

Changes to Existing Functions

The Image Processing Toolbox, Version 3.2, includes changes to these existing functions.

Function

Description of Change

circshift

Moved into MATLAB

freqz2

Checks for insignificant real part in addition to insignificant imaginary part

getnhood

Returns a logical array

gray2ind

More efficient memory usage

imfill

New syntax for grayscale images does not require 'holes' argument. This option is selected automatically.

imlincomb

Accepts more than two images as input and you can specify the output class

immovie

Flicker during movie creation eliminated

imtransform

Linear and bicubic interpolation are faster

ordfilt2

Uses a different algorithm for binary images that improves processing speed for these images

roifilt2

More efficient. Operation is performed only on the region of interest, not the entire image.

  


 © 1984-2008- The MathWorks, Inc.    -   Site Help   -   Patents   -   Trademarks   -   Privacy Policy   -   Preventing Piracy   -   RSS