Model Zoo support for OpenVino 2020.3.2 LTS

[Tye__Stephen]

I have notice new models being made available for OpenVino 2021.3 such as 

text-recognition-0013

Publish second set of candidate models for 2021.3

Will these models also be made available for the LTS version 2020.3.2? Or do they already support it in this format?

Thanks.

[Vladimir_Dudnik]

Unfortunately no, due to new features, implemented in OpenVINO since last 2020.3 LTS release, new models IR not compatible with old 2020.3 runtime. Is there any specific reason to not migrate to the latest available OpenVINO release?

[Vladimir_Dudnik]

An additional note, I'd like to note that in OpenVINO 2021.4 which is coming soon, we do plan to replace text-recognition-0013 model with improved text-recognition-0014 model. Also, there will be a new text-recognition-0015 model, built on different topology. Please stay tuned!

[Tye__Stephen]

Thank you. We are planning to upgrade to OpenVino 2021.4 when it comes out. 

Will either 0014 or 0015 support the OCR of characters like : / . 

We wish to identify URLs in an image but currently we cannot output the :// or the . with the existing model.

[Vladimir_Dudnik]

@Tye__Stephen  you can look at text-detection-0014 and text-detection-0015 models description on Open Model Zoo github, just choose develop branch (default is the master branch which corresponds to the latest public release)

I'll let development team know about your request.

Also, please note that text spotting-0005 (text detection and recognition) model available in OpenVINO Training Extensions, so you may re-train it for your particular task.

Regards,
  Vladimir

[Munesh_Intel]

Hi Stephen,

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

Regards,

Munesh