DeepLabv3+ and OpenVINO

[paularamo]

Hi @PINTO0309 ! Thanks for this amazing repo. Congrats!.

I'm looking for a way to run DeepLabv3+ with OpenVINO. I see that in your repository you have several folders related to this model. It seems that the one that best fits my needs is Sample.4 - Semantic Segmentation, DeeplabV3-plus 256x256. But I would like to do the development simply by calling the executable files and the XML, as in the typical OpenVINO examples. This example requires TensorFlow installation. What about OpenVINO-bin? I did not find a lot of documentation about this latter one.
I am using RPi4 and a OAK-D camera.

Thanks for your help.

[PINTO0309]

Please wait a little because I will be home in 16 hours.

[PINTO0309]

Sorry. I guess I'm too busy today to get any sleep. I'll be releasing it later, so please be patient for a bit.

[PINTO0309]

@pjramg
I have committed a sample program. Please feel free to use it.
https://github.com/PINTO0309/PINTO_model_zoo/tree/master/26_mobile-deeplabv3-plus/07_openvino

FP16 + CPU 30 FPS

FP16 + USB3.0 + MYRIAD 15 FPS

[paularamo]

That sounds great. I'll be home in a day. I'll let you know how it went with your model. Thank you very much.

[paularamo]

Hi @PINTO0309. I am trying to run your code, I am not sure what is the issue about. So, I share my screenshot. Please let me some lights to continue. Thanks.

[PINTO0309]

Follow the official tutorial to update to the latest OpenVINO 2020.4.
https://docs.openvinotoolkit.org/latest/openvino_docs_install_guides_installing_openvino_raspbian.html

[paularamo]

Thanks, @PINTO0309. It was the version I installed on RPi.

[paularamo]

Some updates on my side. I am trying to fix this issue, and follow a the error that terminal shows me. So, it seems that it is because of I only have the OAK-D camera plug to RPi, and it is not recognize directly, the RPi recognizes the VPU. Also I select this device with --device MYRIAD using python3 deeplabv3plus_usbcam.py --device MYRIAD and I got this warning. [ WARN:0] global ../opencv/modules/videoio/src/cap_v4l.cpp (893) open VIDEOIO(V4L2:/dev/video0): can't open camera by index. Let me know if you have some ideas I could use. Thanks for your valuable help.

[PINTO0309]

Check what port the USB camera or the built-in camera is recognized on. You just need to specify the recognized port number as the program argument. --usbcamno The default is 0.

$ ls -l /dev/video*

[paularamo]

Maybe I have a problem with the cameras (BW1098FFC).

I realized that they don't support the board enumerated as a webcam (usb UVC protocol)

[PINTO0309]

I don't own an OAK-D so I can't try it. Follow the official tutorial below and give it a try.
https://github.com/luxonis/depthai-tutorials.git

[paularamo]

The OAk-D camera is running well Object detection models, but not Semantic Segmentation models. The examples from thy webpage are running well, I am trying to run deeplav3+ on the camera as a NCS2 (OpenVINO) and follow your examples.

[PINTO0309]

Please modify my sample program to access the camera in a different way.
https://github.com/luxonis/depthai-tutorials/blob/0f4e3feadad4003dc153eac06a2e80d3009b2de8/2-face-detection-retail/face-detection-retail-0004.py#L5-L32

[PINTO0309]

There is no progress, so I'll close it.