The Vitis Video Analytics SDK (VVAS) is a framework to build transcoding and AI-powered solutions on Xilinx platforms. It takes input data - from USB/CSI camera, video from file or streams over RTSP, and uses Vitis™ AI to generate insights from pixels for various use cases. VVAS SDK can be the foundation layer for a number of video analytic solutions like understanding traffic and pedestrians in a smart city, health and safety monitoring in hospitals, self-checkout and analytics in retail, detecting component defects at a manufacturing facility and others. VVAS can also be used to build Adaptive Bitrate Transcoding solutions that may require re-encoding the incoming video at different bitrates, resolution, and encoding format.
The core SDK consists of several hardware accelerator plugins that use various accelerators such as multiscaler (for resize and color space conversion), and deep learning processing unit (DPU) for machine learning. By performing all the compute heavy operations in dedicated accelerators, VVAS can achieve highest performance for video analytics, transcoding and several other application areas.
- Ships several hardware accelerators for various functions
- Provides highly optimized GStreamer plugins, which meets most of the requirements of the Video Analytics and transcoding solutions
- Provides an easy to use framework to integrate the hardware accelerators/kernels in Gstreamer framework based applications
- Provides AI model support for popular object detection and classification models like SSD, YOLO etc
- Application developers can build seamless streaming pipelines for AI-based video and image analytics, complex Adaptive Bitrate Transcoding pipelines, and several other solutions using VVAS, without having any understanding about low level environment complexities.
- VVAS provides the flexibility for rapid prototyping to full production level solutions by significantly reducing the time to market for the solutions on Xilinx platforms.