Getting Access to the PVA with VPA

⟵ ⟶

RidgeRun NVIDIA PVA Development

Table of Contents [Sticky]

• Introduction
  • What is the PVA?
  • When to Use the PVA?
  • Development Workflow
  • Services
• Getting Access to the PVA
  • Vision Programming Interface(VPI)
  • GStreamer VPI Plug-In
  • Custom Applications
• Algorithms
• Contact Us
• Sponsor your Favorite Feature

Vision Programming Interface (VPI)

A possible approach to get access to the power of the PVA is through the NVIDIA Vision Programming Interface (VPI), which is a cross-platform API that allows developers to write vision processing code abstracted from the underlying hardware. Supported backends include CPU, GPU, VIC, and PVA.

Using VPI, developers can:

• Rapidly prototype vision algorithms.
• Switch between hardware backends without rewriting application logic.
• Seamlessly transition algorithms to the PVA for efficient execution.

PVA support in VPI includes operations such as:

• Box and Gaussian filtering
• Morphological operations
• Laplacian and Gaussian pyramids
• Image remapping and warping
• Dense optical flow estimation

The VPI framework handles memory transfers, backend scheduling, and task queuing, reducing development complexity. This makes it particularly useful for applications that may need to fall back on CPU or GPU when PVA is unavailable. It is important to mention that VPI is a library from NVIDIA and the code is not available for extensions.

For more information, please, visit the VPI website.

⟵ ⟶