mm011what - Syntax, data types and conventions.

Synopsis

mm011what

Description

What is the SDC Morphology Toolbox

The SDC Morphology Toolbox is a set of functions (i.e., programs) for Image and Signal Processing. The functions available in our toolbox are based on the theory of Mathematical Morphology, a powerful non linear approach for the representation of image and signal transforms.

Our toolbox is useful for solving several Image and Signal Processing problems, particularly, segmentation, extraction of quantitative information, noise reduction and compression.

An interesting characteristic of Mathematical Morphology is that it permits to study generic transformations of discrete signals directly (i.e., without appealing to continuous approximations). This fact permits the development of efficient algorithms on compact integer data structures, that is a remarkable characteristic of our software.

Platforms and implementation

The SDC Morphology Toolbox runs on MATLAB and is supported for 3 platforms: MS-Windows, Linux and Solaris.

The functions of our toolbox are implemented as mex files. For each function, there is an m-file and a mex-file, that is, a .dll, a .mexlx or a .mexsol, depending on the platform. The mex-files call a dynamic library, implemented in C and named libmorph.dll or libmorph.so, depending on the platform.

Syntax conventions

The SDC Morphology Toolbox is generated mostly using automatic methods. This ensures that the syntax and conventions used are very consistent throughtout the toolbox.

The toolbox operates on the following data structures:

• Images (1-D, 2-D, or 3D arrays) with pixels of types uint8, uint16, and logical uint8 (binary)
• Structuring elements
• Interval
• Strings
• Scalar values as double
• Boolean values

General rules

• Each function of the toolbox can return only one variable
• The operands on the toolbox can be optional only from left to right. See mmconv - Conventions
• The most common default parameter is the structuring element mmsecross (elementare 3x3 cross)
• All the images in most functions must have the same data type and same size
• You can operate an image with a constant value. The constant is regarding as a constant image with same size as the other
• The use of mmfreedom aliviates this requirement, controlling the automatic data type conversions

Use of mmfreedom

• Can have values 0, 1 or 2:
• 0: does not allow any automatic conversion
• 1: makes automatic conversions and send a warning message
• 2: makes automatic conversions quietly
• The default value is 1
• We strongly recommend to switch to value 0, at the beginning. This can help explain many unexpected results

Gray-scale and Binary images

• The gray-scale images are of type uint8 (0-255) and uint16 (0-65535)
• The binary images are of type logical uint8 and has only the values 0 and 1
• There are two functions to convert from binary to gray-scale images: mmgray and mmbinary

Exercises

• What is the main difference of an image with zero and ones stored as uint8 and stored as logical uint8. Think of a problem that may happen if the two images are used in the wrong situation.
• See mmdataconv and explain the examples of automatic data type conversions.

[Top] [Up] [Prev] [Next]	[Up] [Tutorial]
Copyright (c) 1998-20001 by SDC Information SystemsCopyright (c) 1998-20001 by SDC Information Systems