SysCV / sam-hq

Segment Anything in High Quality [NeurIPS 2023]
https://arxiv.org/abs/2306.01567
Apache License 2.0
3.7k stars 223 forks source link

A very good SAM fine-tuning job #1

Open LXiu-yu opened 1 year ago

LXiu-yu commented 1 year ago

This is an exciting work, I read your article, I would like to ask if your team has tested it on remote sensing imagery, which is a relatively complex task.

lkeab commented 1 year ago

Hi, thanks for watching our work. We didn't try on the remote sensing images yet.

LXiu-yu commented 1 year ago

Hi, thanks for watching our work. We didn't try on the remote sensing images yet.

Thank you for your reply and very much look forward to your work being open sourced in the community. I will try to put your research to the test in our study if I get the chance!

lkeab commented 1 year ago

We will release it within June. Please stay tuned!

zzcgithub commented 1 year ago

respect

KNIND commented 1 year ago

A very good improvement over the 'Foundation Model". Could this be extended for point clouds also in the future?

giswqs commented 1 year ago

I am adding support for HQ-SAM in the segment-geospatial package, which is designed for segmenting remote sensing image. See https://github.com/opengeos/segment-geospatial/pull/161