Closed argman closed 6 years ago
Hi,
Thanks for your attention.
We haven't done it yet, but we are working on it. Currently our main concern is for curve text recognition based on our dataset. We will also test its universality in our extended version.
Thanks for sharing your great idea, but seems the paper only have results on CTW, what about the performance on other datasets