search

yikaiw / CEN

[TPAMI 2023, NeurIPS 2020] Code release for "Deep Multimodal Fusion by Channel Exchanging"
MIT License
284 stars 43 forks source link

Some question about the image size #4

Closed counter-clz closed 3 years ago

counter-clz commented 3 years ago

Thanks for your excellent work！I have some question about the input size of NYUDv2, Why the processed image size is not 480 x 640 in provided NYUDv2 dataset? and what is the AlignToMask() transformation used for in NYUDv2 and SUNRGBD dataset?

Thank you very much and look forward your reply.

yikaiw commented 3 years ago

Thank you for your interest. The provided NYUDv2 dataset is processed in https://github.com/DrSleep/light-weight-refinenet, which crops white borders of RGB and mask images. As this repo does not deal with the depth data, the function AlignToMask is for cropping the depth images, in order to match them with RGB or mask images. In other datasets, AlignToMask is actually not needed, because RGB, depth, and the mask already have the same shape.

counter-clz commented 3 years ago

Thank you for your quickly reply