Spaces:
Running
on
Zero
Running
on
Zero
MaskAdapter
Mask-Adapter
Mask-Adapter: The Devil is in the Masks for Open-Vocabulary Segmentation
YongKang Li1,*, Tianheng Cheng1,*, Wenyu Liu1, Xinggang Wang1,π§
1 Huazhong University of Science and Technology,
(* equal contribution, π§ corresponding author)
Highlights
Mask-Adapter is a simple yet remarkably effective method and can be seamlessly integrated into open-vocabulary segmentation methods, e.g., FC-CLIP and MAFT-Plus, to tackle the existing bottlenecks.
Mask-Adapter effectively extends to SAM without training, achieving impressive results across multiple open-vocabulary segmentation benchmarks.
Updates
- Release code
- Release weights
- Release demo with SAM-2π π€ Mask-Adapter
- Release weights training with addtional data
Installation
Please follow installation.
Getting Started
See Preparing Datasets for Mask-Adapter.Following FC-CLIP and MAFT-Plus to prepare datasets.
See Getting Started with Mask-Adapter.
Citing Mask-Adapter
If you use Mask-Adapter in your research, please use the following BibTeX entry.