This repo provides the source code for "Cross-Domain Adaptive Teacher for Object Detection".

Overview

Cross-Domain Adaptive Teacher for Object Detection

License: CC BY-NC 4.0

License: CC BY-NC 4.0

This is the PyTorch implementation of our paper:
Cross-Domain Adaptive Teacher for Object Detection
Yu-Jhe Li, Xiaoliang Dai, Chih-Yao Ma, Yen-Cheng Liu, Kan Chen, Bichen Wu, Zijian He, Kris Kitani, Peter Vajda
IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2022

[Paper] [Project]

Installation

Prerequisites

  • Python ≥ 3.6
  • PyTorch ≥ 1.5 and torchvision that matches the PyTorch installation.
  • Detectron2 == 0.3 (The version I used to run my code)

Install python env

To install required dependencies on the virtual environment of the python (e.g., virtualenv for python3), please run the following command at the root of this code:

$ python3 -m venv /path/to/new/virtual/environment/.
$ source /path/to/new/virtual/environment/bin/activate

For example:

$ mkdir python_env
$ python3 -m venv python_env/
$ source python_env/bin/activate

Build Detectron2 from Source

Follow the INSTALL.md to install Detectron2.

Dataset download

  1. Download the datasets

  2. Organize the dataset as the Cityscapes and PASCAL VOC format following:

adaptive_teacher/
└── datasets/
    └── cityscapes/
        ├── gtFine/
            ├── train/
            └── test/
            └── val/
        ├── leftImg8bit/
            ├── train/
            └── test/
            └── val/
   └── cityscapes_foggy/
        ├── gtFine/
            ├── train/
            └── test/
            └── val/
        ├── leftImg8bit/
            ├── train/
            └── test/
            └── val/
   └── VOC2012/
        ├── Annotations/
        ├── ImageSets/
        └── JPEGImages/
   └── clipark/
        ├── Annotations/
        ├── ImageSets/
        └── JPEGImages/
   └── watercolor/
        ├── Annotations/
        ├── ImageSets/
        └── JPEGImages/
    

Training

  • Train the Adaptive Teacher under PASCAL VOC (source) and Clipart1k (target)
python train_net.py \
      --num-gpus 8 \
      --config configs/faster_rcnn_R101_cross_clipart.yaml\
  • Train the Adaptive Teacher under cityscapes (source) and foggy cityscapes (target)
python train_net.py\
      --num-gpus 8\
      --config configs/faster_rcnn_VGG_cross_city.yaml\
      OUTPUT_DIR output/exp_city

Resume the training

python train_net.py \
      --resume \
      --num-gpus 8 \
      --config configs/faster_rcnn_R101_cross_clipart.yaml MODEL.WEIGHTS <your weight>.pth

Evaluation

python train_net.py \
      --eval-only \
      --num-gpus 8 \
      --config configs/faster_rcnn_R101_cross_clipart.yaml \
      MODEL.WEIGHTS <your weight>.pth

Results and Model Weights

Real to Artistic Adaptation:

Backbone Source set (labeled) Target set (unlabeled) Batch size [email protected] Model Weights Comment
R101 VOC12 Clipark1k 16 labeled + 16 unlabeled 40.6 link (coming soon) Ours w/o discriminator
R101 VOC12 Clipark1k 16 labeled + 16 unlabeled 49.3 link (coming soon) Ours in the paper
R101+FPN VOC12 Clipark1k 16 labeled + 16 unlabeled 51.2 link (coming soon) For future work

Weather Adaptation:

Backbone Source set (labeled) Target set (unlabeled) Batch size [email protected] Model Weights Comment
VGG16 Cityscapes Foggy Cityscapes 16 labeled + 16 unlabeled 48.7 link (coming soon) Ours w/o discriminator
VGG16 Cityscapes Foggy Cityscapes 16 labeled + 16 unlabeled 50.9 link (coming soon) Ours in the paper
VGG16+FPN Cityscapes Foggy Cityscapes 16 labeled + 16 unlabeled 57.4 link (coming soon) For future work

Citation

If you use Adaptive Teacher in your research or wish to refer to the results published in the paper, please use the following BibTeX entry.

@inproceedings{li2022cross,
    title={Cross-Domain Adaptive Teacher for Object Detection},
    author={Li, Yu-Jhe and Dai, Xiaoliang and Ma, Chih-Yao and Liu, Yen-Cheng and Chen, Kan and Wu, Bichen and He, Zijian and Kitani, Kris and Vajda, Peter},
    booktitle={IEEE Conference on Computer Vision and Pattern Recognition (CVPR)},
    year={2022}
} 

Also, if you use Detectron2 in your research, please use the following BibTeX entry.

@misc{wu2019detectron2,
  author =       {Yuxin Wu and Alexander Kirillov and Francisco Massa and
                  Wan-Yen Lo and Ross Girshick},
  title =        {Detectron2},
  howpublished = {\url{https://github.com/facebookresearch/detectron2}},
  year =         {2019}
}

License

This project is licensed under CC-BY-NC 4.0 License, as found in the LICENSE file.

Owner
Meta Research
Meta Research
Automates the process to obtain an appointment for NIE in spain.

get-nie-appointment A Python script that automates the process of getting an appointment for NIE assignation. It can be modified in order to change th

Ezequiel Aceto 39 Nov 28, 2022
A multi-tenant multi-client scalable product categorising demo stack

Better Categories 4All: A multi-tenant multi-client product categorising stack The steps to reproduce training and inference are in the end of this fi

7 Feb 15, 2022
A simple Python wrapper for the archive.is capturing service

archiveis A simple Python wrapper for the archive.is capturing service. Installation pipenv install archiveis Python Usage Import it. import archi

PastPages 157 Dec 28, 2022
Social Framework

Social Int Framework Social Int Framework its a Selenium script that scrape the IG photos and do a Reverse search on google and yandex for finding ano

29 Dec 06, 2022
Auto Join: A GitHub action script to automatically invite everyone to the organization who comment at the issue page.

Auto Invite To Org By Issue Comment A GitHub action script to automatically invite everyone to the organization who comment at the issue page. What is

Max Base 6 Jun 08, 2022
A telegram bot that can upload telegram media files to anonfiles.com and give you direct download link

✯ AnonFilesBot ✯ Telegram Files to AnonFiles Upload Bot It will Also Give Direct Download Link Process : Fork This Repositry And Simply Cick On Heroku

Avishkar Patil 38 Dec 30, 2022
The official command-line client for spyse.com

Spyse CLI The official command-line client for spyse.com. NOTE: This tool is currently in the early stage beta and shouldn't be used in production. Yo

Spyse 43 Dec 08, 2022
A Python library for the Discourse API

pydiscourse A Python library for working with Discourse. This is a fork of the original Tindie version. It was forked to include fixes, additional fun

Ben Lopatin 72 Oct 14, 2022
A tool for creating credentials for accessing S3 buckets

s3-credentials A tool for creating credentials for accessing S3 buckets For project background, see s3-credentials: a tool for creating credentials fo

Simon Willison 138 Jan 06, 2023
KaydyPurge - Python Purge Script for Discord made by Kaydy Cain#0001

How to Install Open terminal Execute "git clone https://github.com/apolo1337/Kay

apolo 5 Jan 27, 2022
Automation that uses Github Actions, Google Drive API, YouTube Data API and youtube-dl together to feed BackJam app with new music

Automation that uses Github Actions, Google Drive API, YouTube Data API and youtube-dl together to feed BackJam app with new music

Antônio Oliveira 1 Nov 21, 2021
Wanna play on the Overwatch NA servers?

OverwatchRegionSwapper-NA- Wanna play on the Overwatch NA servers? Use at you own risk. LIST OF IPs EU Netherlands: From: 5.42.168.0 to: 5.42.175.255

1 Jun 08, 2022
Demonstrate how GitHub OIDC token getting should be included in boto3

boto3 should add direct support for AssumeRoleWithWebIdentity for GitHub Actions There is a aws-actions/configure-aws-credentials action that will get

Ben Kehoe 11 Aug 29, 2022
A repository for 8G server's discord bot

8G Discord-Bot A general-purpose discord bot for the 8G Discord-Server To setup: Create a new file called secrets.py and make it look like this TOKEN=

1 Jan 12, 2022
Automation application was made by me using Google, Sheet and Slack APIs with Python.

README This application is used to transfer the data in the xlsx document we have to the Google Drive environment and calculate the "total budget" wit

3 Apr 12, 2022
A Python Discord bot project generator

Heater Heat up a Discord bot in a blink What is Heater? Heater is a Command Line Interface tool which allows you to generate a barebones Python Discor

DevGuyAhnaf 5 Jan 14, 2022
A collection of automation aids to connect various database systems into Lookout for Metrics

A collection of automation aids to connect various database systems into Lookout for Metrics

AWS Samples 3 Apr 28, 2022
Unofficial Medium Python Flask API and SDK

PyMedium - Unofficial Medium API PyMedium is an unofficial Medium API written in python flask. It provides developers to access to user, post list and

Engine Bai 157 Nov 11, 2022
Get some python in google cloud functions

[NOTE]: This is a highly experimental (and proof of concept) library so do not expect all python packages to work flawlessly. Also, cloud functions ar

Martin Abelson Sahlen 200 Nov 24, 2022
A simple Discord Mass-Ban that's still working with Member Scraper.

Mass-Ban [!] This was made for education / you can use for revenge. Please don't skid it. [!] If you want to use it, please use member scraper before

WoahThatsHot 1 Nov 20, 2021