Face Detection OpenVX* Workload Sample

This Face Detection OpenVX* sample demos how to use OpenVX* in your application. It implements Intel private face detection algorithm based on classifier cascade detection.

Face detection algorithm is chosen for illustrative purposes, real production face detection pipelines would involve more advanced algorithms to obtain potentially more robust results. These algorithms are not covered here.

This sample teaches you how to write complete OpenVX* application and also how to develop an example of workload library which can be connected with higher layer framework, e.g., Gstreamer. It introduces advanced features required for a real application. If you rather need a detailed step-by-step introduction to the basics of OpenVX* development, see Auto Contrast sample, available in this SDK (<SDK_ROOT>/samples/auto_contrast).

The following topics are covered in the sample:

  • Expressing user-defined logic as a part of a graph via user nodes.
  • Using of Intel experimental feature called Targets API for heterogeneous computing.
  • Obtaining and interpreting OpenVX* performance information.
  • How to create workload library to connect with upper layer framework, e.g., Gstreamer
  • How to do face detection perf optimization

Additionally, the sample features basic interoperability with OpenCV through data sharing. OpenCV is used for reading the data from a video file or camera device.

For more complete information about compiler optimizations, see our Optimization Notice.