search

fabiopoiesi / dip

Project page of the paper "Distinctive 3D local deep descriptors" accepted in IEEE International Conference on Pattern Recognition 2020.
64 stars 4 forks source link

Questions about data generation #17

Open qsisi opened 2 years ago

qsisi commented 2 years ago

Thanks for your sharing, here I have a question about how to generate 3DMatch_train.zip data.

As far as I know, the data downloaded from 3DMatch contains 62 RGB-D sequences, do those *.ply files were generated by TSDF fusion provided by 3DMatch-toolbox?

Thanks for your help.

fabiopoiesi commented 2 years ago

Correct, this is the tool. In my case I used the ones provided by this paper, I didn't recompute them.

qsisi commented 2 years ago

Thanks for your reply, I could not correctly use the tool provided by 3DMatch due to the opencv configuration :(

It seems that this info: "3DMatch_train/7-scenes-fire/cloud_bin_19.info.txt" is missing in 3DMatch_train.zip ?

qsisi commented 2 years ago

I got a new question :)

The 3DMatch_train.zip contains 82 folders, which in my understanding, includes the whole training scenes in 3DMatch, according to the split.txt provided in the original 3DMatch website, there are total 54 scenes for training.

So it means, 54 scenes of rgbd-sequences are converted into 82 folders of *.ply files? (54 scenes -> 82 folders)

But when I check the fuseSceneFragments.m, it seems that it just fuse each scene into one folder, so maybe 54 scenes -> 54 folders?

So why the 3DMatch_train.zip contains 82 folders? Maybe they are not generated using the 3DMatch-toolbox?

Thank you very much for your help!