Intel® Distribution of OpenVINO™ Toolkit
Community support and discussions about the Intel® Distribution of OpenVINO™ toolkit, OpenCV, and all things computer vision-related on Intel® platforms.
6059 Discussions

NEW RELEASE: 'facial-landmarks-98-detection-0001' heatmap output

Anzhela_G
Beginner
376 Views

Hi,

I wanted to test the new landmarks-detection model on my software (https://docs.openvino.ai/latest/omz_models_model_facial_landmarks_98_detection_0001.html) 
as I use Openvino Toolkit for its development.

The output of this model is a 1x98x16x16 matrix, which is described  as location heatmaps for 98 points in the documentation. I have some trouble with handling this output as I've done a research about how a location heatmap could be parsed into (x,y) coordinates to be displayed on an image and didn't found any way to make a progress. I also researched about superimposing a heatmap on the base image but this model gives a regions for specific points not the heatmap of a whole image containing landmark points as I understood.

Could you give me any ideas about how could I handle this output, please.

Regards.

0 Kudos
1 Solution
Peh_Intel
Moderator
334 Views

Hi Anzhela,


Thanks for reaching out to us.


The facial-landmarks-98-detection-0001 model can be used in Gaze Estimation Demo.


You can refer to the landmarks_estimator.cpp and landmarks_estimator.hpp.



Regards,

Peh


View solution in original post

4 Replies
Peh_Intel
Moderator
335 Views

Hi Anzhela,


Thanks for reaching out to us.


The facial-landmarks-98-detection-0001 model can be used in Gaze Estimation Demo.


You can refer to the landmarks_estimator.cpp and landmarks_estimator.hpp.



Regards,

Peh


Anzhela_G
Beginner
317 Views

Hi Peh,

Thanks for your quick reply.

 

Now I am facing VPU(MYRIAD) detection issue on my device, as after my latest installation of new version and doing all required configurations mentioned in documentation it doesn't show me MYRIAD available after running ./INSTAL_DIR/samples/python/hello_query_device/hello_query_device.py script.

 

After first run of the script it crashes with Runtime error.
After second run it shows CPU, GPU, GNA and HDDL without MYRIAD.

 

Could you help me out with this issue, please. I am also attaching pictures of terminal messages of the script. 

 

Regards,

Anzhela

Peh_Intel
Moderator
283 Views

Hi Anzhela,


VPU devices include Intel® Neural Compute Stick (supported by MYRIAD Plugin) and

Intel® Vision Accelerator Design with Intel® Movidius™ VPUs (supported by the HDDL Plugin). However, MYRIAD plugin is unable to run inferencing with HDDL plugin together. Hence, MYRIAD is not being detected after running the HDDL Plugin backend service (hddldaemon).


You can run inferencing with HDDL plugin instead of MYRIAD plugin as you have Intel® Vision Accelerator Design with Intel® Movidius™ VPUs on your host machine.


If you wish to run inferencing with MYRIAD plugin, kill the HDDL Plugin backend service (hddldaemon) and reset all Intel® Movidius™ VPUs before running an application that uses the MYRIAD Plugin:

kill -9 $(pidof hddldaemon autoboot)

pidof hddldaemon autoboot # Make sure none of them is alive

source /opt/intel/openvino_2021/bin/setupvars.sh

${HDDL_INSTALL_DIR}/bin/bsl_reset



Regards,

Peh


Peh_Intel
Moderator
230 Views

Hi Anzhela,


This thread will no longer be monitored since we have provided answers. If you need any additional information from Intel, please submit a new question. 



Regards,

Peh


Reply