search

hkchengrex / Mask-Propagation

[CVPR 2021] MiVOS - Mask Propagation module. Reproduced STM (and better) with training code :star2:. Semi-supervised video object segmentation evaluation.
https://hkchengrex.github.io/MiVOS/
MIT License
127 stars 22 forks source link

thin-plate-spline question #37

Closed cheind closed 2 years ago

cheind commented 2 years ago

Hey,

I'm the author of https://github.com/cheind/py-thin-plate-spline and discovered your dependency here. I've tracked its usage down to https://github.com/hkchengrex/Mask-Propagation/blob/b5d8e61c87e1944426c3aed43685fd0eb336353a/dataset/tps.py#L16

I'm curious, what are you using it for? It seems like you are augmenting training data by warping the data with thin-plate-spline model. Was the library useful in this respect, i.e did it improve performance (quantifyable)? I'd like add a linkto your paper/model the project page if you don't mind?

hkchengrex commented 2 years ago

Hi,

Thank you for your code. I am using it for data augmentation. I have not specifically checked the quantitative improvement from this augmentation but the generated results do look promising. I liked how it is able to augment the image beyond standard affine/perspective transformation.
TPS was used in a previous paper: https://arxiv.org/pdf/1612.02646.pdf so I just followed from there.

I have also used it in my other paper: https://github.com/hkchengrex/STCN for the same purpose.

You are welcomed to add links.

cheind commented 2 years ago

Hey,

thanks for the information! I'll have a look at the papers mentioned.