Specifications
Intel
®
Image Processing Library Reference Manual
2-2
2
The multi-spectral image (MSI) model is used for general purpose images.
It is used for any kind of multi-spectral data and any kind of image. For
example, the Fourier transform operation writes transform coefficients of
color or monochrome images to this modelone channel for each channel
in the input. The result can be viewed as an MSI image. An MSI image can
contain any number of color channels; they may even correspond to
invisible parts of the spectrum. The library functions do not need to
identify any specific MSI image channels.
Data Types and Palettes
The parameter that determines the image data type is the pixel depth in
bits. The data could be signed integer, unsigned integer, or floating-point.
The following data types are supported for various color models
(s = signed, u = unsigned, f = float):
Gray scale 1, 8s, 8u, 16s, 16u, and 32f bits per pixel
Color (three-channel) 8u and 16u bits per channel
Four-channel and MSI 8s, 8u, 16s, 16u, 32s, and 32f bits per channel.
The library supports only absolute color images in which each pixel is
represented by the channel intensities. For example, in an absolute color
24-bit RGB image, three bytes (24 bits) per pixel represent the three
channel intensities. LUT (lookup table) images, that is, palette color images
are not supported. You must convert palette images to absolute color
images for further processing by the library functions. There are special
functions for converting DIB palette images to absolute color images.
Color images with 8, 16, or 32 bits per channel simply pack each channel,
respectively, into a byte, word, or doubleword. All channels within a given
image have the same data type.
Signed data (8s, 16s, or 32s) are used for storing the output of some image
processing operations; for example, this is the case for transforms such as
FFT. Unless specified otherwise, signed data cannot be used as input to
image processing operations.