deep-hough-transform icon indicating copy to clipboard operation
deep-hough-transform copied to clipboard

Why parametric space label is generated?

Open HassanBinHaroon opened this issue 3 years ago • 3 comments

When we download the dataset, it contains the ground truth.

Kindly enlighten is it necessary to create parametric space label while creating a custom dataset?

HassanBinHaroon avatar Aug 09 '22 09:08 HassanBinHaroon

If you want to train DHT on your custom dataset, the straight line annotations is needed to calculate the parameteric space lable. Then train your model with the parameteric space label.

Hanqer avatar Aug 10 '22 10:08 Hanqer

If I create the straight line annotations then How can I calculate parametric space label?

There are two scripts of calculating parametric space label each for different dataset (prepare_data_NKL.py & prepare_data_JTLEE.py).

Can any of these two (prepare_data_NKL.py & prepare_data_JTLEE.py) be used If I create my custom dataset to create parametric space label?

HassanBinHaroon avatar Aug 10 '22 11:08 HassanBinHaroon

@Hanqer @zeakey @garipovroma Kindly take a look at my Github repository. I tried to reproduce your work on Google Colab. Sometimes its convenient to run on Google Colab for the users and for instant results visualization. If you approve the work then you can provide reference of my repository on yours. Thanks!

https://github.com/HassanBinHaroon/deep-hough-transform

HassanBinHaroon avatar Aug 11 '22 02:08 HassanBinHaroon