MAFT+ (ECCV 2024 oral)

Collaborative Vision-Text Representation Optimizing for Open-Vocabulary Segmentation

Siyu Jiao^1,2,Hongguang Zhu^1,2,Jiannan Huang^1,3, Yao Zhao^1,2, Yunchao Wei^1,2, Humphrey Shi^3,4,

¹ Beijing Jiaotong University, ² Pengcheng Lab, ³ Georgia Institute of Technology, ⁴ Picsart AI Research (PAIR)

[Paper]

Introduction

This work is an enhanced version of our NeurIPS paper MAFT.
Pre-trained vision-language models, e.g. CLIP, have been increasingly used to address the challenging Open-Vocabulary Segmentation (OVS) task, benefiting from their well-aligned vision-text embedding space. Typical solutions involve either freezing CLIP during training to unilaterally maintain its zero-shot capability, or fine-tuning CLIP vision encoder to achieve perceptual sensitivity to local regions. However, few of them incorporate vision-text collaborative optimization. Based on this, we propose the Content-Dependent Transfer to adaptively enhance each text embedding by interacting with the input image, which presents a parameter-efficient way to optimize the text representation. Besides, we additionally introduce a Representation Compensation strategy, reviewing the original CLIP-V representation as compensation to maintain the zero-shot capability of CLIP. In this way, the vision and text representation of CLIP are optimized collaboratively, enhancing the alignment of the vision-text feature space. To the best of our knowledge, we are the first to establish the collaborative vision-text optimizing mechanism within the OVS field. Extensive experiments demonstrate our method achieves superior performance on popular OVS benchmarks. In open-vocabulary semantic segmentation, our method outperforms the previous state-of-the-art approaches by +0.5, +2.3, +3.4, +0.4 and +1.1 mIoU, respectively on A-847, A-150, PC-459, PC-59 and PAS-20. Furthermore, in a panoptic setting on ADE20K, we achieve the performance of 27.1 PQ, 73.5 SQ, and 32.9 RQ.

Installation

1. Clone the repository

   git clone https://github.com/jiaosiyu1999/MAFT_Plus.git
   

2. Navigate to the project directory

   cd MAFT_Plus
   

3. Install the dependencies

   bash install.sh
   cd maft/modeling/pixel_decoder/ops
   sh make.sh
   

Data Preparation

See MAFT for reference (Preparing Datasets for MAFT). The data should be organized like:

datasets/
  ade/
      ADEChallengeData2016/
        images/
        annotations_detectron2/
      ADE20K_2021_17_01/
        images/
        annotations_detectron2/
  coco/
        train2017/
        val2017/
        stuffthingmaps_detectron2/
  VOCdevkit/
     VOC2012/
        images_detectron2/
        annotations_ovs/      
    VOC2010/
        images/
        annotations_detectron2_ovs/
            pc59_val/
            pc459_val/      

Usage

• Pretrained Weights

  1. semantic

  Model	A-847	A-150	PC-459	PC-59	PAS-20	Weights
  MAFTP-Base	13.8	34.6	16.2	57.5	95.4	maftp_b.pth
  MAFTP-Large	15.1	36.1	21.6	59.4	96.5	maftp_l.pth

  2. panoptic

  	PQ	SQ	RQ	Weights
  MAFTP-Large	27.1	73.5	32.9	maftp_l_pano.pth

• Evaluation

  evaluate trained model on validation sets of all datasets.

  python train_net.py --eval-only --config-file <CONFIG_FILE> --num-gpus <NUM_GPU> OUTPUT_DIR <OUTPUT_PATH> MODEL.WEIGHTS <TRAINED_MODEL_PATH>
  

  For example, evaluate our pre-trained maftp_l.pth model:

  # 1. Download MAFTP-Large.
  # 2. put it at `out/semantic/MAFT_Plus/maftp_l.pth`.
  # 3. evaluation
    python train_net.py --config-file configs/semantic/eval.yaml  --num-gpus 8 --eval-only \
                         MODEL.WEIGHTS out/semantic/MAFT_Plus/maftp_l.pth 
  

• Training

  end to end training requires 8*A100 GPUs and 14 hours, approximately:

    # MAFT-Plus-Large (maftp-l)
    python train_net.py --config-file configs/semantic/train_semantic_large.yaml  --num-gpus 8

    # MAFT-Plus-Base (maftp-b)
    python train_net.py --config-file configs/semantic/train_semantic_base.yaml  --num-gpus 8

• Inference Demo with Pre-trained Models

  We provide demo/demo.py that is able to demo builtin configs. Run it with:

  python demo/demo.py \
    --input input1.jpg input2.jpg \
    [--other-options]
    --opts MODEL.WEIGHTS /path/to/checkpoint_file
  

  For example, evaluate our pre-trained maftp_l.pth model:

  # 1. Download MAFTP-Large.
  # 2. put it at `out/semantic/MAFT_Plus/maftp_l.pth`.
  # 3. run demo:
    python demo/demo.py  --input im.png
  

Cite

If this codebase is useful to you, please consider citing:

@inproceedings{jiao2024collaborative,
  title={Collaborative Vision-Text Representation Optimizing for Open-Vocabulary Segmentation},
  author={Jiao, Siyu and Zhu, Hongguang and Huang, Jiannan and Zhao, Yao and Wei, Yunchao and Humphrey, Shi},
  booktitle={European Conference on Computer Vision},
  year={2024},
}

Acknowledgement

Mask2Former

FC-CLIP