Kai Li

more details on using the code

1

nice work and thanks for making it public. Could you please provide more introduction about how to use the code, including the environment, the dataset, the access to the datasets....

Could you provide the details on preparing the datasets? It seems not the organized in the way that was adopted as the previous methods, e.g., SW-DET? Thanks.

number of test images

4

Hi, the provided dataset has 347 images for testing, while the paper says that there are 500 images for testing. I wonder why there is such a difference. Thanks.

upload code for trinet

1

Hi, I think your idea is inspiring and would like to build on your work. Could you upload the trinet part of your algorithm some time? Your generosity is much...

Hi, could you upload the code for the other datasets. Thanks.

mse loss is way much higher than the kl loss in eq (3), how do you deal with the imbalance?

1

Hi I am trying to reimplement the code. My current implementation shows that in eq (3), the mse loss is way much higher than the kl loss in eq (3)...

Lunet (train from scatch)

4

Nice work. I wonder to know if it is possible to release the implementation of the proposed LuNet? Thanks.

does the regularization loss make a lot difference on the results?

1

nice work. does the regularization loss make a lot difference on the results? Is there a big gap with it removed? Thanks.

Great work! May I ask what are the hardware requirement and training time? Thanks a lot.