Skip to content

Vstream - Video Analytics pipeline with Hardware based accelerations (dev - stage)

Notifications You must be signed in to change notification settings

VimalWill/Vstream

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

84 Commits
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Vstream - Video Analytics Pipeline

The pipeline is constructed on the top of Neural-Net Inference Engine called "Infera" which currently supports YoloV8 object detection model and Streaming pipeline buiild on the base of GStreamer framework. The Infera supports google-test for the functionality verification of each functions in the system.

Requirements

  • OpenCV (above 4.8-dev)
  • OpenCL
  • GStreamer
  • Google-Test for C++
  • Standard C++ 17

Build Procedure for Vstream

The building process for Vstream handles building both Test-suit for Infera and Final Executable for the Vstream
$ mkdir build && cd build
$ cmake ..
$ make -j64
  • The exectuables will be created for both Global Test-suit and Vstream

Features of Vstream

  • Adapative Hardware Offloading of Hardware for Inference
  • Hardware Accelerated Video Compression for network stream (on-progress)

GIF Example

To-Dos

  • Hardware acceleration for video encoding
  • Network Streaming
  • Object Tracking

About

Vstream - Video Analytics pipeline with Hardware based accelerations (dev - stage)

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published