• 04/03/2020
  • Public Content

OpenVX* Objects

This chapter provides a brief description of some essential OpenVX* objects to understand basic OpenVX* application development. Refer to the OpenVX* 1.1 specification for more details.
OpenVX context is a container for OpenVX* objects. You have to create a context to be able to create objects. A program can have more than one context for various reasons. However, sharing objects across contexts is very limited.
OpenVX kernel is the abstract representation of a computer vision function, such as a “Sobel Gradient” or “Lucas Kanade Feature Tracking.” Kernels are referred by qualified names (e.g.
) or enumerations, and the functionality is defined by the specification and conformance tests.
OpenVX node is an instance of a kernel. The node encapsulates the kernel (functionality) and the position of that functionality in the data flow of the graph. The latter is defined by assigning the node parameters, for example, when output of one node becomes an input for another node. Every node that is created has to belong to a graph.
OpenVX graph is the collection of the nodes with their connections. The graph must be directed (only goes one-way) and acyclic (does not loop back). OpenVX* supports two modes of graph execution, synchronous/blocking mode (using
) and asynchronous (via
). In both modes, the order of execution of nodes inside a single graph is defined by the data-flow which is guaranteed. However, in asynchronous mode there is no defined order for graph execution (each specified via
call). To control the order of execution, you have to insert
Graphs in OpenVX* depend on data objects to link together nodes. OpenVX image is an opaque data object that holds pixels. There are multiple types of images. For example, to enable certain optimizations, the intermediate images can be declared as virtual. Such images cannot be accessed for read or write outside the graph, as there is no associated accessible memory. Basically, virtual images are stubs to define a dependency on OpenVX nodes (just like temporary variables in C/C++), and these can be eliminated by the OpenVX graph compiler.

Product and Performance Information


Intel's compilers may or may not optimize to the same degree for non-Intel microprocessors for optimizations that are not unique to Intel microprocessors. These optimizations include SSE2, SSE3, and SSSE3 instruction sets and other optimizations. Intel does not guarantee the availability, functionality, or effectiveness of any optimization on microprocessors not manufactured by Intel. Microprocessor-dependent optimizations in this product are intended for use with Intel microprocessors. Certain optimizations not specific to Intel microarchitecture are reserved for Intel microprocessors. Please refer to the applicable product User and Reference Guides for more information regarding the specific instruction sets covered by this notice.

Notice revision #20110804