Intel® Distribution of OpenVINO™ Toolkit
Community assistance about the Intel® Distribution of OpenVINO™ toolkit, OpenCV, and all aspects of computer vision-related on Intel® platforms.

Combine two YOLO models

Benjamin1
Novice
2,378 Views

Hi, I am trying to put together two YOLO models to detect different objects. I would like to know if there is a code(in Python)  to do this easily, I am using the object_detection_demo_yolov3_async.py, but it only works for a single model (for example the generic YOLO for 80 classes).
I would also like to know if a YOLO model can be combined with any of the pre-training models from the intel page.

Thanks a lot

0 Kudos
5 Replies
Iffa_Intel
Moderator
2,366 Views

Greetings,


This is an example of how to use multiple models in one application:

https://software.intel.com/content/www/us/en/develop/videos/part-6-multiple-models-in-one-application.html


Sincerely,

Iffa


0 Kudos
Benjamin1
Novice
2,361 Views

Hi, thanks :), I had already seen that video, however it does not explain how to do it with python.

Regards

 

0 Kudos
Iffa_Intel
Moderator
2,347 Views

That actually depends on which OpenVINO application you are running. It's good if you apply -h to it to see what they support. For instance, the Pedestrian Tracker demo uses two IR models.

 

You may refer here:

https://docs.openvinotoolkit.org/2018_R5/_samples_pedestrian_tracker_demo_README.html

This should apply to any demos for any language. Again don't forget to use -h command.

Sincerely,

Iffa

 

0 Kudos
Benjamin1
Novice
2,340 Views

Hi , thanks , which I have to use for two models that detect different objects (for example 2 different yolo models).

I did not find any script that has those requirements.

0 Kudos
Iffa_Intel
Moderator
2,331 Views

Actually, if you are just trying to differentiate between two objects you don't need to use two models.

 

Instead, train the model to differentiate the objects like here:

https://www.youtube.com/watch?v=FaqVhvJ6-Uc

OR

https://www.youtube.com/watch?v=cbdS3BjjbaQ

 

 

However, if you are trying to detect text and audio OR detect face, age and gender, then you need to use different models like here: 

https://www.youtube.com/watch?v=HEntm0TUqM8

 

Sincerely,

Iffa

 

 

Reply