Home

Awesome

<br /> <p align="center"> <h1 align="center">Spider: A Unified Framework for Context-dependent Concept Segmentation</h1> <p align="center"> ICML, 2024 <br /> <a href="https://xiaoqi-zhao-dlut.github.io/"><strong>Xiaoqi Zhao*</strong></a> · <a href="https://lartpang.github.io/"><strong>Youwei Pang*</strong></a> . <a href="https://jiwei0921.github.io/"><strong>Wei Ji*</strong></a> . <a ><strong>Baicheng Sheng</strong></a> . <a ><strong>Jiaming Zuo</strong></a> · <a href="https://scholar.google.com/citations?hl=zh-CN&user=XGPdQbIAAAAJ"><strong>Lihe Zhang*</strong></a> · <a href="https://scholar.google.com/citations?hl=zh-CN&user=D3nE0agAAAAJ"><strong>Huchuan Lu</strong></a> </p> <p align="center"> <a href='https://arxiv.org/pdf/2405.01002'> <img src='https://img.shields.io/badge/Paper-PDF-green?style=flat&logo=arXiv&logoColor=green' alt='arXiv PDF'> </a> </p> <br />

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

<p align="center"> <img src="./image/CI_vs_CD.png"/> <br /> </p>

CD Concept Segmentation Survey Paper

(IJCV 2024) Towards Diverse Binary Segmentation via A Simple yet General Gated Network

Unified 8 CD Concept Segmentation Tasks

<p align="center"> <img src="./image/UniverCDSeg.png"/> <br /> </p>

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

<p align="center"> <img src="./image/Spider.png"/> <br /> </p>

Performance

<p align="center"> <img src="./image/performance1.png"/> <br /> </p> <p align="center"> <img src="./image/performance2.png"/> <br /> </p> <p align="center"> <img src="./image/performance3.png"/> <br /> </p> <p align="center"> <img src="./image/performance4.png"/> <br /> </p>

Potential

Continual/Zero-shot/Incremental Zero-shot learning

<p align="center"> <img src="./image/ZSL1.png"/> <br /> </p> <p align="center"> <img src="./image/ZSL2.png"/> <br /> </p>

In-Context Learning

<p align="center"> <img src="./image/In_context_learning.png"/> <br /> </p>

Datasets

<p align="center"> <img src="./image/datasets.png"/> <br /> </p>

Trained Models

Prediction Maps

To Do List

Citation

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

@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}