> For the complete documentation index, see [llms.txt](https://amobile-solutions.gitbook.io/yocto-developer-guide/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://amobile-solutions.gitbook.io/yocto-developer-guide/test-and-use-the-interface/nnstreamer.md).
# NNStreamer
{% hint style="info" %}
For more details, please visit the official MediaTek IoT Yocto [NNStreamer page](https://mediatek.gitlab.io/aiot/doc/aiot-dev-guide/master/tools/nnstreamer.html).
{% endhint %}
{% hint style="info" %}
The following instructions utilize the SD510 for demonstration purposes. These procedures are applicable to other development boards.
{% endhint %}
## NNStreamer Unit Test
NNStreamer offers a [gtest-based test suite](https://github.com/nnstreamer/nnstreamer/tree/main/tests) for both its common library and plugins. To assess NNStreamer's integration status on IoT Yocto, execute the unit tests with the following command.
```sh
cd /usr/bin/unittest-nnstreamer/
ssat
```
Here takes ArmNN unit test as an example.
```sh
cd /usr/bin/unittest-nnstreamer/tests/
export NNSTREAMER_SOURCE_ROOT_PATH=/usr/bin/unittest-nnstreamer/
./unittest_filter_armnn
```
## NNStreamer Examples
Explore the Python examples available in `/usr/bin/nnstreamer-demo/` on IoT Yocto to learn how to build NNStreamer pipelines with varied `tensor_filters` configurations for different applications. These examples are adapted from the [NNStreamer-Example](https://github.com/nnstreamer/nnstreamer-example) repository.
| Category | Input Source |
| -------------------------- | ------------ |
| Image Classification | Camera |
| Object Detection | Camera |
| Object Detection | Camera |
| Pose Estimation | Camera |
| Face Detection | Camera |
| Monocular Depth Estimation | Camera |
| Image Enhancement | Image |
### General Configuration
{% hint style="info" %}
The SBC-MB510 MIPI-CSI camera is still in development. Please use a device with UVC camera functionality to run the following example.
{% endhint %}
#### Get camera node ID
To obtain the correct **camera node ID** with the following command:
```sh
v4l2-ctl --list-devices
```
We've provided the following example:
```
root@sd510:~# v4l2-ctl --list-devices
...
Integrated Camera: Integrated C (usb-11200000.xhci1-1.2):
/dev/video5
/dev/video6
/dev/media1
...
```
In this case, the camera node ID is **5** and Type is **uvc**
### Demo Runner
Usage
```
usage: run_nnstreamer_example.py [-h] [--app {image_classification,object_detection,object_detection_yolov5,face_detection,pose_estimation,low_light_image_enhancement,monocular_depth_estimation}]
[--engine {neuronsdk,tflite,armnn}] [--img IMG] [--cam CAM] --cam_type {uvc,yuvsensor,rawsensor} [--width WIDTH] [--height HEIGHT] [--performance {0,1}] [--fullscreen {0,1}] [--throughput {0,1}]
[--rot ROT]
options:
-h, --help show this help message and exit
--app {image_classification,object_detection,object_detection_yolov5,face_detection,pose_estimation,low_light_image_enhancement,monocular_depth_estimation}
Choose a demo app to run. Default: image_classification
--engine {neuronsdk,tflite,armnn}
Choose a backends to inference. Default: neuronsdk
--img IMG Input a image file path.
Example: /usr/bin/nnstreamer-demo/original.png
Note: This paramater is dedicated to low light enhancement app
--cam CAM Input a camera node id, ex: 130 .
Use 'v4l2-ctl --list-devices' query camera node id.
Example:
$ v4l2-ctl --list-devices
...
C922 Pro Stream Webcam (usb-11290000.xhci-1.2):
/dev/video130
/dev/video131
...
Note: This paramater is for all the apps except low light enhancement app.
--cam_type {uvc,yuvsensor,rawsensor}
Choose correct type of camera being used for the demo, ex: yuvsensor
Note: This paramater is for all the apps except low light enhancement app.
--width WIDTH Input video display width, ex: 640
--height HEIGHT Input video display height, ex: 480
--performance {0,1} Enable to make CPU/GPU/APU run under performance mode, ex: 1
--fullscreen {0,1} Fullscreen preview.
1: Enable
0: Disable
Note: This paramater is for all the apps except low light enhancement app and monocular depth estimation app.
--throughput {0,1} Print throughput information.
1: Enable
0: Disable
--rot ROT Rotate the camera image by degrees, ex: 90
Note: This paramater is for all the apps except low light enhancement app.
```
### Examples
{% hint style="info" %}
The following example engine is neuronsdk and performance mode is enabled. \
For other engines and parameters, please refer to the official website.
{% endhint %}
#### Image Classification
Example command:
```
root@sd510:/usr/bin/nnstreamer-demo# python3 run_nnstreamer_example.py --app image_classification --engine neuronsdk --cam 5 --cam_type uvc --performance 1 --fullscreen 1
--app image_classification --engine neuronsdk --cam 5 --cam_type uvc --performance 1 --fullscreen 1
INFO:root:finished to load labels, total [1001]
INFO:root:Set performance mode
INFO:root:pipeline: v4l2src name=src device=/dev/video5 io-mode=mmap ! video/x-raw,width=640,height=480,format=YUY2 ! tee name=t_raw t_raw. ! queue leaky=2 max-size-buffers=10 ! textoverlay name=label font-desc=Sans,18 valignment=top ! fpsdisplaysink name=sink text-overlay=false signal-fps-measurements=true sync=false video-sink="waylandsink sync=false qos=false fullscreen=1"t_raw. ! queue leaky=2 max-size-buffers=2 ! videoconvert ! videoscale ! video/x-raw,width=224,height=224,format=RGB ! tensor_converter ! tensor_filter framework=neuronsdk throughput=0 name=nn model=/usr/bin/nnstreamer-demo/mobilenet_v1_1.0_224_quant.dla inputtype=uint8 input=3:224:224:1 outputtype=uint8 output=1001:1 ! tensor_sink name=tensor_sink
INFO:root:Run: Image classification.
INFO:root:received start message
```
