search

UX-Decoder / Semantic-SAM

[ECCV 2024] Official implementation of the paper "Semantic-SAM: Segment and Recognize Anything at Any Granularity"
2.36k stars 116 forks source link

When will the code and demo with semantics be released? #41

Open becauseofAI opened 1 year ago

becauseofAI commented 1 year ago

If I want to know the categories of segmentation (semantic, instance, part), can this code do it? If so, what should I do to achieve this goal?

FengLi-ust commented 1 year ago

We have not released the code with semantic labels. Please stay tuned for this code and checkpoint!

becauseofAI commented 1 year ago

We have not released the code with semantic labels. Please stay tuned for this code and checkpoint!

When will the code with semantic labels be released? I can't wait anymore. @FengLi-ust

geweihgg commented 1 year ago

+1, can't wait anymore ^-^

chongruo commented 1 year ago

Could we have the expected timeline? thanks

shenkev commented 1 year ago

My research project is also trying to use the semantic labels so it'd be nice if that can be released soon, otherwise we will need to try other methods instead...

AhmedBourouis commented 1 year ago

+1 waiting for the code with semantic labels

BellalaLv commented 10 months ago

My research project is also trying to use the semantic labels so it'd be nice if that can be released soon, otherwise we will need to try other methods instead...

Are there any other semantic models you can recommend? They still haven't released the code。

brucejunlee commented 9 months ago

This is a great work! When will the code with semantic labels be released, please? soon? There are some open-source projects which have implemented this module with semantic labels, such as Grounded-SAM (+SAM-HQ), X-Decoder, and OpenSeeD (built on Mask DINO), but these models have worse performance in classes and boundaries than Semantic-SAM from the viewpoint of visualization.