VisionWorks Toolkit Reference

December 18, 2015 | 1.2 Release

 All Data Structures Namespaces Files Functions Variables Typedefs Enumerations Enumerator Macros Groups Pages
VisionWorks Toolkit Reference Documentation

Welcome to the NVIDIA® VisionWorks™ Toolkit API reference. VisionWorks implements and extends the Khronos OpenVX standard.

With the VisionWorks library of Computer Vision primitives, and its framework optimized for NVIDIA platforms, you can develop mobile and embedded applications for Robotics, Drones, Autonomous Driving, Augmented Reality (AR), Intelligent Video Analytics, and other Computer Vision markets.

With recent innovations in mobile and embedded technologies and the proliferation of visual sensors with increased computational power in small form factors, NVIDIA leads the way to enabling development of applications targeting these innovative, new modalities, which enhance interactions between the real and the virtual worlds. Mobile and embedded devices are typically battery-powered, requiring performance and power efficiency for fluid user experiences.

NVIDIA® VisionWorks™ Toolkit provides a production-quality, standard interface that implements and extends the OpenVX standard, to enable application development for high performance on NVIDIA mobile and embedded devices. Where applicable, VisionWorks technologies take advantage of underlying hardware acceleration, such as NVIDIA® CUDA® technology, enabling significant speedup and thereby reducing engineering efforts in designing real-time vision applications.

VisionWorks has been optimized with a focus on mobile and embedded platforms that use NVIDIA® Tegra® processors, featuring multi-core ARM CPU architecture and a CUDA-enabled embedded GPU. Support for traditional desktop configurations with discrete GPU and x86-64 CPU architecture are also provided for ease of development; however, VisionWorks performance has not been tuned for desktop platforms.

Contents

Documentation is preliminary and subject to change. These references include the following sections:

  • Architecture — Provides toolkit architecture, theory of operation information, and an overview of primitives and algorithms.
  • Programming Model Overview — Summarizes the most important aspects of the programming model, which takes the OpenVX specification model as its base.
  • Installation Guide — Provides information to help you get started using VisionWorks.
  • NVIDIA Extension API — Provides detailed VisionWorks API usage information. For more information, see:
    • Framework APIs — Provides NVIDIA Framework Extension API usage information.
    • Vision Primitives APIs — Provides NVIDIA Vision Extension API usage information.
    • NVXIO APIs — Sample library to demonstrate input/output functionality that you can use in your applications.
  • Khronos OpenVX 1.0.1 API — Provides OpenVX specification.
  • Samples and Demos — Provides usage information for demos and samples that demonstrate the specific functionality.
  • Tutorials — Demonstrates how to use VisionWorks framework with CUDA and how to create custom nodes. This section includes full source code of the tutorials.

The following illustration shows examples of applications that can benefit from VisionWorks.

nvx_mainpage_pic.png