CNStream is a streaming framework with plug-ins. It is used to connect other modules, includes basic functionalities, libraries, and essential elements.
CNStream provides the following built-in modules:
- DataSource: Support RTSP, video file, images, elementary stream in memory and sensor inputs (H.264, H.265, and JPEG decoding) (sensor input is only supported on edge platforms).
- Inferencer: MLU-based inference accelerator for detection and classification, based on EasyDK InferServer.
- Osd (On-screen display): Module for highlighting objects and text overlay.
- VEncode: Encode videos or images and write to file or push RTSP stream to internet.
- Vout: Display the video on screen (Only support on edge platforms).
- Tracker: Multi-object tracking.
To start using CNStream, please refer to the chapter of quick start in the document of Cambricon-CNStream-User-Guide-CN.pdf .
| Classification | Object Detection |
|---|---|
 |
 |
| Object Tracking | License plate recognition |
|---|---|
 |
 |
| Body Pose |
|---|
 |
Modify the files.list_video file, which is under the samples directory, to replace the video path. Each line represents one stream. It is recommended to use an absolute path or use a relative path relative to the executor path.
Check out the Examples page for tutorials on how to use CNStream. Concepts page for basic definitions.
Discuss - General community discussion around CNStream.
