A practical and incremental approach to learn OpenVX
This project presents a series of simple examples that aim to introduce basic OpenVX concepts. It was developed as a companion to the OpenVX Training presentation.
In order to build the project you'll need the following dependencies:
- Makefile, C and C++ compilers: In Debian based systems (like Ubuntu) you may run the following commands:
sudo apt install build-essential- OpenCV
In Debian based systems (like Ubuntu) you may run the following commands:
sudo apt install libopencv-dev- A working OpenVX implementation:
Normally this is already included in your system. If not, you may install the reference implmenentation from Khronos.
The project is based on a simple Makefile. To build it, run the following commands:
makeIf your OpenVX installation is on a non-standard location, you may pass in the respective flags as in:
make VX_CFLAGS="-I/non/standard/vx/includes" VX_LDFLAGS="-L/non/standard/vx/lib"To run the examples, simply invoke the executables produced during the build process as in:
./vx_training_01Of course, the 00 may be changed to match any example as desired. Refer to the following sections for a description of each example.
The following table summarizes the examples available in the project. They were numbered to, ideally, be consumed in order.
| Example | Description | 1st Arg | 2nd Arg |
|---|---|---|---|
| vx_training_01 | Creates a context and shows the overall VX development pattern. | ||
| vx_training_02 | Creates an image and shows how to access the underlying memory in order to write to it. | Image path (defaults to lena.png) | |
| vx_training_03 | Creates a graph with a Channel Extract node and an output image. Does ot process the graph yet. | Image path (defaults to lena.png) | |
| vx_training_04 | Verifies and executes the graph. Saves the data from the output image into a PNG file. | Image path (defaults to lena.png) | Image path (defaults to out.png) |
| vx_training_05 | Adds a second Gaussian Kernel node and connects it to the first one using a virtual image. | Image path (defaults to lena.png) | Image path (defaults to out.png) |
| vx_training_06 | Adds a third Warp Affine node and shows how to pass in a vx_reference as a parameter. | Image path (defaults to lena.png) | Image path (defaults to out.png) |
| vx_training_07 | First example in C++. Shows how to continuously process the graph and vary a parameter with each execution. Displays the result in a window. | Image path (defaults to lena.png) | |
| vx_training_08 | Shows how to enable performance measurements. | Image path (defaults to lena.png) | |
| vx_training_09 | Modifies the previous example to be executed in a pipelining mode. | Image path (defaults to lena.png) | |
| vx_training_10 | Modifies the previous example to be executed in a batching mode. | Image path (defaults to lena.png) |
If you run into any problem or have any question, please do contact us.