Licence-plate-recognition-barrier-0001

[Gdeep]

Intel model   'Vehicle-license-plate-detection-barrier' is only working on few colour of number plates like yellow, blue,

and intel Model 'licence-plate-recognition-barrier-001' is only recognized Chinese number plates

This both Intel_models are not working on Indian plates , so please tell me any way or which steps I need to be done so it also work with Indian plates...

Best regards

Gagandeep

[Severine_H_Intel]

Dear Gagan, 

for the vehicle license plate detector, you can use this new tool we have released that enables you to retrain it: https://github.com/opencv/training_toolbox_tensorflow

For the second one, it is not yet possible to retrain it. 

Best, 

Severine

[Gdeep]

Dear Habert,

They mentioned steps for Linux but am using open-vino on Window plate-form ? Can this process also same with window plate-form?

And My second question is:-

1 Download training data and put it in the data/bitvehicle directory according to data/bitvehicle/README.md file. Annotation files in the COCO format (refer to cocodataset for details) are already located in data/bitvehicle.

From where to I could download training data??

Best regards

Gagandeep

[Severine_H_Intel]

Dear Gagandeep, 

unfortunately, the tool is only for Ubuntu for now. For the second question,  I am not sure I understood well the question, but you can find the training data here: http://iitlab.bit.edu.cn/mcislab/vehicledb/ However, if you want to retrain for the Indian plates and vehicles, it would be better if you have your own Indian specific dataset. 

Best, 

Severine

[Gdeep]

Dear Habert

Can we except this tool also released for Window plate-form in future 

best regards

Gagandeep

[Severine_H_Intel]

Dear Gagan, I dont think this is the priority for now. 

Best, 

Severine