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.
5770 Discussions

OpenVINO model zoo: camvid semantic segmentation

Letty
Novice
538 Views

Hi,

I would like to try and use some of the models available in the OpenVINO model zoo that use the Camvid dataset (e.g. icnet-camvid-ava-0001).

However I am confused by the description of the output, as it says that the classes used are 12 (https://github.com/openvinotoolkit/open_model_zoo/blob/master/models/intel/icnet-camvid-ava-0001/des...), but the Camvid dataset on which it is based has 32 classes instead (http://mi.eng.cam.ac.uk/research/projects/VideoRec/CamVid/).

Is it possible to know which classes are the output of the icnet-camvid-ava-0001 model and in what order are the class IDs?

Labels (1)
0 Kudos
1 Solution
IntelSupport
Community Manager
504 Views

Hi Letty,

The icnet-camvid-ava-0001 model is based on 12 classes version used for the training. Check out this research paper at section 3.2 and 4 on the CamVid and it classes available.

https://arxiv.org/pdf/1608.01745.pdf

 

Regards,

Aznie


View solution in original post

4 Replies
IntelSupport
Community Manager
511 Views

Hi Letty,

Thanks for reaching out. We are investigating this and will get back to you at the earliest.


Regards,

Aznie


IntelSupport
Community Manager
505 Views

Hi Letty,

The icnet-camvid-ava-0001 model is based on 12 classes version used for the training. Check out this research paper at section 3.2 and 4 on the CamVid and it classes available.

https://arxiv.org/pdf/1608.01745.pdf

 

Regards,

Aznie


Letty
Novice
500 Views

Thanks a lot for the reply and for the link to the original paper, appreciate your help!

Regards

IntelSupport
Community Manager
487 Views

Hi Letty,

This thread will no longer be monitored since this issue has been resolved. If you need any additional information from Intel, please submit a new question.


Regards,

Aznie




Reply