CSIRO Mathematics, Informatics and Statistics
|
|
|
Next: Image analysis routine design Up: What is a LIAR Previous: What is a LIAR
The Image HeaderAll image analysis systems written in C use some kind of image header data structure for holding image information such as: image type, image size, the image data (i.e. pixels), image coordinates and pointers to other data structures and functions. For example, Fig. 2 provides a very basic image header, for raster scanned images. Data structures for image processing are a popular because they help to produce a common interface (common coupling) to the underlying modules of an image analysis system. If there was just one image analysis system then this simple interface approach would be all that is needed. The problem is, there are many image analysis systems with their own common interface mechanisms.
It follows, that different IAS's use data structures that essentially contain the same type of basic information. In an attempt to standardise the interface between low level functions and an IAS, a standard set of image data structures could be used. To cope with existing IAS's, input image data would be first mapped to the standard data structure(s) before being passed to the lower level routines. However, this approach is fraught with difficulties, since the standard data structure must be comprehensive enough to allow for specialized requirements in a host of scientific areas now and in the future. Therefore, the LIAR methodology is based on an alternative more elegant and succinct approach.
Ed Breen Tue Sep 3 17:19:31 EST 1996 |