Intel® Integrated Performance Primitives Developer Guide and Reference

ID 790148
Date 11/07/2023
Public

A newer version of this document is available. Customers should click here to go to the newest version.

Document Table of Contents

Glossary

absolute colors

Colors specified by each pixel's coordinates in a color space. Intel Integrated Performance Primitives for image processing use images with absolute colors.

alpha channel

A color channel, also known as the opacity channel, that can be used in color models; for example, the RGBA model.

arithmetic operation

An operation that adds, subtracts, multiplies, divides, or squares the image pixel values.

color-twist matrix

A matrix used to multiply the pixel components in one color space for determining the components in another color space.

DCT

Acronym for the discrete cosine transform. SeeDiscrete Cosine Transforms in “Image Linear Transforms”.

dilation

A morphological operation that sets each output pixel to the minimum of the corresponding input pixel and its 8 neighbors.

dyadic operation

An operation that has two input images. It can have other input parameters as well.

element-wise operation

An element-wise operation performs the same operation on each element of a vector, or uses the elements of the same position in multiple vectors as inputs to the operation.

erosion

A morphological operation that sets each output pixel to the maximum of the corresponding input pixel and its 8 neighbors.

four-channel model

A color model that uses four color channels; for example, the RGBA color model.

gray scale image

An image characterized by a single intensity channel so that each intensity value corresponds to a certain shade of gray.

in-place operation

An operation whose output image is one of the input images.

linear filtering

In this document, 2D convolution operations.

linear image transforms

In this document, the discrete cosine transform (DCT).

MMX™ technology

An enhancement to the Intel® architecture aimed at better performance in multimedia and communications applications. The technology uses four additional data types, eight 64-bit MMX registers, and 57 additional instructions implementing the SIMD (single instruction, multiple data) technique.

monadic operation

An operation that has a single input image. It can have other input parameters as well.

morphological operation

An erosion, dilation, or their combinations.

not-in-place operation

An operation whose output is an image other than the input image(s). See in-place operation.

pixel depth

The number of bits determining each channel intensity for a single pixel in the image.

pixel-oriented ordering

Storing the image information in such an order that the values of all color channels for each pixel are clustered; for example, RGBRGB... .

planar-oriented ordering

Storing the image information so that all data of one color channel follow all data of another channel, thus forming a separate “plane” for each channel; for example, RRRRRGGGGGBBBBB....

region of interest

A rectangular image region on which an operation acts (or processing occurs).

RGB

Red-green-blue. A three-channel color model that uses red, green, and blue color channels.

RGBA

Red-green-blue-alpha. A four-channel color model that uses red, green, blue, and alpha (or opacity) channels.

ROI

See identity matrix.

row-major order

The default storage method for arrays in C. Memory representation is such that the rows of an array are stored contiguously. For example, for the array a[3][4], the element a[1][0] immediately follows a[0][3].

saturation

Using saturation arithmetic, when a number exceeds the data-range limit for its data type, it saturates to the upper data-range limit. For example, a signed word greater than 7FFFh saturates to 7FFFh. When a number is less than the lower data-range limit, it saturates to the lower data-range. For example, a signed word less than 8000h saturates to 8000h.

Intel® Streaming SIMD Extensions

The enhancement to the Intel architecture instruction set for the next generation processors. It incorporates a group of general-purpose floating-point instructions operating on packed data, additional packed integer instructions, together with cacheability control and state management instructions. These instructions significantly improve performance of applications using compute-intensive processing of floating-point and integer data.

three-channel model

A color model that uses three color channels; for example, the RGB color model.