Panoptic Scene Graph Generation

<img src="./assets/psgtr_long.gif" align="center" width="80%"> <a href="https://arxiv.org/abs/2207.11247" target='_blank'> <img src="https://img.shields.io/badge/Paper-ECCV%202022-b31b1b?style=flat-square"> </a>     <a href="https://psgdataset.org/" target='_blank'> <img src="https://img.shields.io/badge/Page-psgdataset.org-228c22?style=flat-square"> </a>     <a href="https://entuedu-my.sharepoint.com/:f:/g/personal/jingkang001_e_ntu_edu_sg/EgQzvsYo3t9BpxgMZ6VHaEMBDAb7v0UgI8iIAExQUJq62Q?e=fIY3zh" target='_blank'> <img src="https://img.shields.io/badge/Data-PSGDataset-334b7f?style=flat-square"> </a>     <a href="https://www.cvmart.net/race/10349/base" target='_blank'> <img src="https://img.shields.io/badge/Competition-PSG Challenge-f2d297?style=flat-square"> </a> <a href="https://huggingface.co/spaces/mmlab-ntu/OpenPSG" target='_blank'> <img src="https://img.shields.io/badge/Demo-HuggingFace-ffca37?style=flat-square"> </a>     <a href="https://paperswithcode.com/task/panoptic-scene-graph-generation/" target='_blank'> <img src="https://img.shields.io/badge/Benchmark-PapersWithCode-00c4c6?style=flat-square"> </a>     <a href="https://join.slack.com/t/psgdataset/shared_invite/zt-1f8wkjfky-~uikum1YA1giLGZphFZdAQ" target='_blank'> <img src="https://img.shields.io/badge/Forum-Slack-4c1448?style=flat-square">     <a href="https://replicate.com/cjwbw/openpsg" target='_blank'> <img src="https://img.shields.io/badge/Replicate-Demo & Cloud API-1b82c2?style=flat-square"> </a> Panoptic Scene Graph Generation <a href="http://jingkang50.github.io/" target='_blank'>Jingkang Yang</a>,  <a href="https://yizhe-ang.github.io/" target='_blank'>Yi Zhe Ang</a>,  <a href="https://www.linkedin.com/in/zujin-guo-652b0417a/" target='_blank'>Zujin Guo</a>,  <a href="https://kaiyangzhou.github.io/" target='_blank'>Kaiyang Zhou</a>,  <a href="http://www.statfe.com/" target='_blank'>Wayne Zhang</a>,  <a href="https://liuziwei7.github.io/" target='_blank'>Ziwei Liu</a> S-Lab, Nanyang Technological University & SenseTime Research

Updates

Oct 31, 2022: We release the full dataset here that we used in the paper. The competition version is here.
Oct 9, 2022: The preliminary round of PSG challenge ends. We will release the entire dataset after the final round starts. Before that, if you want to access the PSG dataset (competition version), please email me.
Sep 4, 2022: We introduce the PSG Classification Task for NTU CE7454 Coursework, as described here.
Aug 21, 2022: We provide guidance on PSG challenge registration here.
Aug 12, 2022: Replicate demo and Cloud API is added, try it here!
Aug 10, 2022: We launched Hugging Face demo 🤗. Try it with your scene!
Aug 5, 2022: The PSG Challenge will be available on International Algorithm Case Competition ! All the data will be available there then! Stay tuned!
July 25, 2022: :boom: We are preparing a PSG competition with ECCV'22 SenseHuman Workshop and International Algorithm Case Competition, starting from Aug 6, with a prize pool of :money_mouth_face: US$150K :money_mouth_face:. Join us on our Slack to stay updated!
July 25, 2022: PSG paper is available on arXiv.
July 3, 2022: PSG is accepted by ECCV'22.

What is PSG Task?

The Panoptic Scene Graph Generation (PSG) Task aims to interpret a complex scene image with a scene graph representation, with each node in the scene graph grounded by its pixel-accurate segmentation mask in the image.

To promote comprehensive scene understanding, we take into account all the content in the image, including "things" and "stuff", to generate the scene graph.

| | |:--:| | PSG Task: To generate a scene graph that is grounded by its panoptic segmentation|

PSG addresses many SGG problems

We believe that the biggest problem of classic scene graph generation (SGG) comes from noisy datasets. Classic scene graph generation datasets adopt a bounding box-based object grounding, which inevitably causes a number of issues:

Coarse localization: bounding boxes cannot reach pixel-level accuracy,
Inability to ground comprehensively: bounding boxes cannot ground backgrounds,
Tendency to provide trivial information: current datasets usually capture frivolous objects like head to form trivial relations like person-has-head, due to too much freedom given during bounding box annotation.
Duplicate groundings: the same object could be grounded by multiple separate bounding boxes.

All of the problems above can be easily addressed by the PSG dataset, which grounds the objects using panoptic segmentation with an appropriate granularity of object categories (adopted from COCO).

In fact, the PSG dataset contains 49k overlapping images from COCO and Visual Genome. In a nutshell, we asked annotators to annotate relations based on COCO panoptic segmentations, i.e., relations are mask-to-mask.

| | |:--:| | Comparison between the classic VG-150 and PSG.|

Clear Predicate Definition

We also find that a good definition of predicates is unfortunately ignored in the previous SGG datasets. To better formulate PSG task, we carefully define 56 predicates for PSG dataset. We try hard to avoid trivial or duplicated relations, and find that the designed 56 predicates are enough to cover the entire PSG dataset (or common everyday scenarios).

Type | Predicates | --- | --- | Positional Relations (6) | over, in front of, beside, on, in, attached to. | Common Object-Object Relations (5) | hanging from, on the back of, falling off, going down, painted on.| Common Actions (31) | walking on, running on, crossing, standing on, lying on, sitting on, leaning on, flying over, jumping over, jumping from, wearing, holding, carrying, looking at, guiding, kissing, eating, drinking, feeding, biting, catching, picking (grabbing), playing with, chasing, climbing, cleaning (washing, brushing), playing, touching, pushing, pulling, opening.| Human Actions (4) | cooking, talking to, throwing (tossing), slicing. Actions in Traffic Scene (4) | driving, riding, parked on, driving on. Actions in Sports Scene (3) | about to hit, kicking, swinging. Interaction between Background (3) | entering, exiting, enclosing (surrounding, warping in)

Get Started

To setup the environment, we use conda to manage our dependencies.

Our developers use CUDA 10.1 to do experiments.

You can specify the appropriate cudatoolkit version to install on your machine in the environment.yml file, and then run the following to create the conda environment:

conda env create -f environment.yml

You shall manually install the following dependencies.

# Install mmcv
## CAUTION: The latest versions of mmcv 1.5.3, mmdet 2.25.0 are not well supported, due to bugs in mmdet.
pip install mmcv-full==1.4.3 -f https://download.openmmlab.com/mmcv/dist/cu101/torch1.7.0/index.html

# Install mmdet
pip install openmim
mim install mmdet==2.20.0

# Install coco panopticapi
pip install git+https://github.com/cocodataset/panopticapi.git

# For visualization
conda install -c conda-forge pycocotools
pip install detectron2==0.5 -f \
  https://dl.fbaipublicfiles.com/detectron2/wheels/cu101/torch1.7/index.html

# If you're using wandb for logging
pip install wandb
wandb login

# If you develop and run openpsg directly, install it from source:
pip install -v -e .
# "-v" means verbose, or more output
# "-e" means installing a project in editable mode,
# thus any local modifications made to the code will take effect without reinstallation.

Datasets and pretrained models are provided. Please unzip the files if necessary.

Before October 2022, we only release part of the PSG data for competition, where part of the test set annotations are wiped out. Users should change the json filename in psg.py (Line 4-5) to a correct filename for training or submission.

**For the PSG competition, we provide psg_train_val.json (45697 training data + 1000 validation data with GT). Participa

OpenPSG

Install / Use

README