Spider-UniCDSeg

KO | EN

gitlite — search

#javascript #react #typescript #go #golang #android #machine-learning #deep-learning #macos #java #awesome-list #pytorch #mcp #openai

Spider-UniCDSeg

(★ 356)

(ICML 2024) Spider: A Unified Framework for Context-dependent Concept Segmentation

파일 탐색기

최종 버전 다운로드 (.zip)

- 1.py
- CI_vs_CD.png
- datasets.png
- In_context_learning.png
- performance1.png
- performance2.png
- performance3.png
- performance4.png
- Spider.png
- UniverCDSeg.png
- ZSL1.png
- ZSL2.png
- DBS_group_prompt.py
- FoldConv.py
- prompt_features.py
- color_mask_img.py
- config.py
- dataset_rgb_strategy2.py
- misc.py
- ssim_loss.py
- utils.py
CITATION.cff
kmeans_choose_prompt.py
prediction_multi_task.py
README.md
run.sh
train.py

// repository documentation

Was this content helpful?

★ 0 (0 ratings)

<h1 align="center">Spider: A Unified Framework for Context-dependent Concept Segmentation</h1>
 
 ICML, 2024
 
 <a href="https://xiaoqi-zhao-dlut.github.io/">Xiaoqi Zhao*</a>
 ·
 <a href="https://lartpang.github.io/">Youwei Pang*</a>
 .
 <a href="https://jiwei0921.github.io/">Wei Ji*</a>
 .
 <a >Baicheng Sheng</a>
 .
 <a >Jiaming Zuo</a>
 ·
 <a href="https://scholar.google.com/citations?hl=zh-CN&user=XGPdQbIAAAAJ">Lihe Zhang*</a>
 ·
 <a href="https://scholar.google.com/citations?hl=zh-CN&user=D3nE0agAAAAJ">Huchuan Lu</a>

## Context-independent (CI) Concept vs. Context-dependent (CD) Concept

 <img src="./image/CI_vs_CD.png"/>

## CD Concept Segmentation Survey Paper
[(IJCV 2024) Towards Diverse Binary Segmentation via A Simple yet General Gated Network](https://arxiv.org/pdf/2303.10396)

## Unified 8 CD Concept Segmentation Tasks

 <img src="./image/UniverCDSeg.png"/>

## Spider: UniCDSeg Framework (You only train and infer once! 100% Unified Parameters.)

 <img src="./image/Spider.png"/>

## Performance

 <img src="./image/performance1.png"/> 
 

 <img src="./image/performance2.png"/> 
 

 <img src="./image/performance3.png"/> 
 

 <img src="./image/performance4.png"/>

## Potential
### Continual/Zero-shot/Incremental Zero-shot learning

 <img src="./image/ZSL1.png"/> 
 

 <img src="./image/ZSL2.png"/>

### In-Context Learning

 <img src="./image/In_context_learning.png"/>

## Datasets 

 <img src="./image/datasets.png"/>

-  **DUTS (SOD)**: [Google Drive](https://drive.google.com/file/d/1S4V05HnVqYYkWCfoki_ALRy9E4AJYhhH/view?usp=sharing)  
-  **COD10K (COD)**: [Google Drive](https://drive.google.com/file/d/10_HwNAmsizkp-266f1tan-VIwNqkOkf9/view?usp=sharing)  
-  **SBU (SD)**: [Google Drive](https://drive.google.com/file/d/1SBUF6dF9aJlPzQGql_9lUSDwM8cN7IG-/view?usp=sharing)  
-  **Trans10K (TOS)**: [Trans10K Website](https://xieenze.github.io/projects/TransLAB/TransLAB.html)  
-  **Five datasets (CPS)**: [Google Drive](https://drive.google.com/file/d/1A29IkVysVPUPy4vu1RklKf4AAD7QvV3x/view?usp=sharing)  
-  **COVID-19 data (COD)**: [Google Drive](https://drive.google.com/file/d/1A29IkVysVPUPy4vu1RklKf4AAD7QvV3x/view?usp=sharing)  
-  **BUSI (BLS)**: [Google Drive](https://drive.google.com/file/d/1A29IkVysVPUPy4vu1RklKf4AAD7QvV3x/view?usp=sharing)  
-  **ISIC18 (SLS)**: [Google Drive](https://drive.google.com/file/d/1A29IkVysVPUPy4vu1RklKf4AAD7QvV3x/view?usp=sharing)
   
## Trained Models
-  Spider-ConvNext-B [Google Drive](https://drive.google.com/file/d/1AepdZrpQh0RvvWz784n1zEfkabZOsfKj/view?usp=sharing)
-  Spider-ConvNext-L [GitHub Release](https://github.com/Xiaoqi-Zhao-DLUT/Spider-UniCDSeg/releases/download/v1.0/Spider_convnext_l.pth)
-  Spider-Swin-B [Google Drive](https://drive.google.com/file/d/1AepdZrpQh0RvvWz784n1zEfkabZOsfKj/view?usp=sharing)
-  Spider-Swin-L [GitHub Release](https://github.com/Xiaoqi-Zhao-DLUT/Spider-UniCDSeg/releases/download/v1.0/Spider_swin_l.pth)

## Prediction Maps
-  Spider-Swin-B [Google Drive](https://drive.google.com/file/d/1dVtJ7as8NXpShGB4QExoooSLl5iGofJx/view?usp=sharing)
-  Spider-Swin-L [Google Drive](https://drive.google.com/file/d/1LcQEHjZap1XEmDT6_QOd6Fi3m9diPoHL/view?usp=sharing)
-  Spider-ConvNeXt-B [Google Drive](https://drive.google.com/file/d/1YJDpcPQk1cMANem0jD5yZpZftImzVwVX/view?usp=sharing)
-  Spider-ConvNeXt-L [Google Drive](https://drive.google.com/file/d/15zUFWKJNNQpNhjnhzlftsDLAXIVaZF8O/view?usp=sharing)

## Evaluation Tools

- <https://github.com/Xiaoqi-Zhao-DLUT/PySegMetric_EvalToolkit>
 
## To Do List

- [x] Release data sets.
- [x] Release model code.
- [x] Release model weights.
- [x] Release model prediction maps.

## Citation

If you think Spider-UniCDSeg codebase are useful for your research, please consider referring us:

```bibtex
@inproceedings{Spider,
  title={Spider: A Unified Framework for Context-dependent Concept Segmentation},
  author={Zhao, Xiaoqi and Pang, Youwei and Ji, Wei and Sheng, Baicheng and Zuo, Jiaming and Zhang, Lihe and Lu, Huchuan},
  booktitle={ICML},
  year={2024}
```

Spider-UniCDSeg

파일 탐색기

오늘의 추천 프로젝트

Was this content helpful?

Recent Feedback

Spider-UniCDSeg

파일 탐색기

오늘의 추천 프로젝트

Was this content helpful?

Recent Feedback

Download README