vignywang / awdesc Goto Github PK

Hello, this is a great job, how is the dataset in your project made?what method is used to select the corresponding images?and what algorithm extracts keypoints and despoints?Can this part of the code work be made public?

Hello, @vignywang -I really appreciate for your work！Congratulations :)
The Figure 14(attention maps) in the paper looks great. Currently I want to try to implement it but I failed. Is it convenient to provide the relevant implementation code?Of course, detailed guidance would be better.
I am always looking forward to your kind response.
Best regards.

Should it be heatmap4 = self.heatmap4(c4) ?

ref: https://github.com/vignywang/AWDesc/blob/711d68395ad7a9557fe93725fbf4566f4ac2fa55/nets/network.py#LL77C31-L77C31

Your work is so attractive!
Can you provide the full training and visualization code? Including the training code and a more detailed readme; the training code of SuperPoint.