Yolov3 Python Demo erroneous detections using Tiny Yolov3

[Tr__Ra]

Using object_detection_demo_yolov3_async.py produces the following output on yolov3 tiny coco model:

 

[Iffa_Intel]

Greetings,

If possible, Could you please help to elaborate more regarding the errors that you are getting?

Sincerely,

Iffa

[Tr__Ra]

The errors are visible in the picture: an extra bounding box for each detection.

[Iffa_Intel · Accepted solution]

 

Greetings,

I tried the model with my own videos and it works fine.

Why don't you try to run with this video sample first and see how it goes:

https://www.pexels.com/video/people-in-the-streets-853993

Here is some insight on how to properly convert yolov3  :

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

And our official documentation:

https://docs.openvinotoolkit.org/latest/openvino_docs_MO_DG_prepare_model_convert_model_tf_specific_Convert_YOLO_From_Tensorflow.html

Sincerely,

Iffa

[Tr__Ra]

Thanks,

Your solution works for yolov3.

[Iffa_Intel]

Greetings,

Intel will no longer monitor this thread since we have provided a solution. If you need any additional information from Intel, please submit a new question

Sincerely,

Iffa