• [arXiv] •
Open-vocabulary semantic segmentation requires models to effectively integrate visual representations with open-vocabulary semantic labels. While Contrastive Language-Image Pre-training (CLIP) models shine in recognizing visual concepts from text, they often struggle with segment coherence due to their limited localization ability. In contrast, Vision Foundation Models (VFMs) excel at acquiring spatially consistent local visual representations, yet they fall short in semantic understanding. This paper introduces ProxyCLIP, an innovative framework designed to harmonize the strengths of both CLIP and VFMs, facilitating enhanced open-vocabulary semantic segmentation. ProxyCLIP leverages the spatial feature correspondence from VFMs as a form of proxy attention to augment CLIP, thereby inheriting the VFMs' robust local consistency and maintaining CLIP's exceptional zero-shot transfer capacity. We propose an adaptive normalization and masking strategy to get the proxy attention from VFMs, allowing for adaptation across different VFMs. Remarkably, as a training-free approach, ProxyCLIP significantly improves the average mean Intersection over Union (mIoU) across eight benchmarks from 40.3 to 44.4, showcasing its exceptional efficacy in bridging the gap between spatial precision and semantic richness for the open-vocabulary segmentation task.
# git clone this repository
git clone https://github.com/mc-lan/ProxyCLIP.git
cd ProxyCLIP
# create new anaconda env
conda create -n ProxyCLIP python=3.10
conda activate ProxyCLIP
# install torch and dependencies
pip install -r requirements.txt
We include the following dataset configurations in this repo:
With background class: PASCAL VOC, PASCAL Context, PASCAL Context 459 (PC459), Cityscapes, ADE20k, ADE847, and COCO-Stuff164k,Without background class: VOC20, Context59 (i.e., PASCAL VOC and PASCAL Context without the background category), and COCO-Object.
For PASCAL Context 459 and ADE847, please follow the CAT-Seg to prepare the datasets. For the other datasets, please follow the MMSeg data preparation document to download and pre-process the datasets. The COCO-Object dataset can be converted from COCO-Stuff164k by executing the following command:
python datasets/cvt_coco_object.py PATH_TO_COCO_STUFF164K -o PATH_TO_COCO164K
python demo.py
Please modify some settings in configs/base_config.py before running the evaluation.
For SAM and MAE, please download the checkpoints from SAM and MAE.
Single-GPU:
python eval.py --config ./config/cfg_DATASET.py --workdir YOUR_WORK_DIR
Multi-GPU:
bash ./dist_test.sh ./config/cfg_DATASET.py
Evaluation on all datasets:
python eval_all.py
Results will be saved in results.xlsx.
@inproceedings{lan2024proxyclip,
title={ProxyCLIP: Proxy Attention Improves CLIP for Open-Vocabulary Segmentation},
author={Mengcheng Lan and Chaofeng Chen and Yiping Ke and Xinjiang Wang and Litong Feng and Wayne Zhang},
booktitle={ECCV},
year={2024},
}
This project is licensed under NTU S-Lab License 1.0. Redistribution and use should follow this license.
This study is supported under the RIE2020 Industry Align- ment Fund – Industry Collaboration Projects (IAF-ICP) Funding Initiative, as well as cash and in-kind contribution from the industry partner(s).
This implementation is based on OpenCLIP and SCLIP. Thanks for the awesome work.
If you have any questions, please feel free to reach out at lanm0002@e.ntu.edu.sg.