HET: Sketching Image Gist

Code for the ECCV 2020 paper: "Sketching Image Gist: Human-Mimetic Hierarchical Scene Graph Generation" (accepted).

The code is partly referred to the project rowanz/neural-motifs and KaihuaTang/VCTree-Scene-Graph-Generation. If you get any problem that cause you unable to run the project, you can check the issues under rowanz/neural-motifs first.

Dependencies

• You may follow these commands to establish the environments under Ubuntu system

Install Anaconda
conda update -n base conda
conda create -n motif pip python=3.6
conda install pytorch=0.3 torchvision cuda90 -c pytorch
bash install_package.sh

Prepare Dataset

Please refer to DATA.md for data preparation.

Set up

1. Update the config file with the dataset paths. Follow the steps in DATA.md.

   • You'll also need to fix your PYTHONPATH: export PYTHONPATH=/home/YourName/ThePathOfYourProject

2. Compile everything. run make in the main directory: this compiles the Bilinear Interpolation operation for the RoIs.

3. Pretrain VG detection. The old version involved pretraining COCO as well, but we got rid of that for simplicity. Run ./scripts/pretrain_detector.sh Note: You might have to modify the learning rate and batch size, particularly if you don't have 3 Titan X GPUs (which is what I used).

   • Download the VG150 pretrained detector checkpoint. You need to change the "-ckpt THE_PATH_OF_INITIAL_CHECKPOINT_MODEL" under ./scripts/train_het.sh
   • Download the VG200 pretrained detector checkpoint (code: fmhf).

How to Train / Evaluation

0. Note that, most of the parameters are under config.py. The training stages and settings are manipulated through ./scripts/train_het.sh Each line of command in train_vctreenet.sh needs to manually indicate "-ckpt" model (initial parameters) and "-save_dir" the path to save model. We list some of our checkpoints here.

   • HetH, PredCls/SgCls, VG150: checkpoint (code: yvt1)
   • HetH, SgDet, VG150: checkpoint (code: n964)
   • HetH-RRM, SgCls, VG200_KR, AAP=area+sal: checkpoint (code: 7i9w)

Other Things You Need To Know

• When you evaluate your model, you will find 3 metrics are printed: 1st, "R@20/50/100" is what we use to report R@20/50/100 in our paper, 2nd, "cls avg" is corresponding mean recall mR@20/50/100 proposed by our paper, "total R" is another way to calculate recall that used in some previous papers/projects, which is quite tricky and unfair, because it almost always get higher recall.
• The tuple match rule and triplet match rule will be applied only if the RRM module is applied.

If this paper/project inspires your work, pls cite our work:

arXiv comming soon. 

• For more information, please visit the homepage: kennethwong.tech