Co-mining: Self-Supervised Learning for Sparsely Annotated Object Detection, AAAI 2021.

MEGVII Research

Last update: Dec 7, 2022

Related tags

Overview

Co-mining: Self-Supervised Learning for Sparsely Annotated Object Detection

This repository is an official implementation of the AAAI 2021 paper Co-mining: Self-Supervised Learning for Sparsely Annotated Object Detection.

Introduction

TL; DR. Co-ming is a self-supervised learning framework for sparsely annotated object detection.

Get Started

Install cvpods following the instructions

# Install cvpods
git clone https://github.com/Megvii-BaseDetection/cvpods.git
cd cvpods 
## build cvpods (requires GPU)
python3 setup.py build develop
## preprare data path
mkdir datasets
ln -s /path/to/your/coco/dataset datasets/coco

Download the sparse-annotations from here of four cases and put them into /coco/annotations/. Note that the annotation of "missing_50p" is from the authors of BRL paper.
For fast evaluation, please download trained model from here.
Run the project

git clone https://github.com/megvii-research/Co-mining.git

# for example(e.g. miss50p)
cd co-mining/retinanet.res101.comining.score.06.miss50p/

# train
pods_train --num-gpus 8

# test
pods_test --num-gpus 8
# test with provided weights
pods_test --num-gpus 8 MODEL.WEIGHTS /path/to/your/model.pth

Results

Model	Multi-scale training	AP (minival)	Link
Comining_RetinaNet_Res50_Full	No	36.8	download
Comining_RetinaNet_Res50_Easy	No	35.4	download
Comining_RetinaNet_Res50_Hard	No	31.8	download
Comining_RetinaNet_Res50_Extreme	No	23.0	download
Comining_RetinaNet_Res101_Miss50p	No	33.9	download

Citing Co-mining

If you find Co-mining useful to your research, please consider citing:

@article{wang2021comining,
  title={Co-mining: Self-Supervised Learning for Sparsely Annotated Object Detection},
  author={Wang, Tiancai and Yang, Tong and Cao, Jiale and Zhang, Xiangyu},
  journal={Proceedings of the AAAI conference on artificial intelligence},
  year={2021}
}

Comments

Working with two stage networks

Hi, Great work! I'm curious if this method works for two stage detectors as well and if it does, what precautions should one take to run it with two stage detectors? thanks

opened by rssaketh 1

This is the repository for the AAAI 21 paper [Contrastive and Generative Graph Convolutional Networks for Graph-based Semi-Supervised Learning].

CG3 This is the repository for the AAAI 21 paper [Contrastive and Generative Graph Convolutional Networks for Graph-based Semi-Supervised Learning]. R

12 Oct 28, 2022

Cross-Image Region Mining with Region Prototypical Network for Weakly Supervised Segmentation

Cross-Image Region Mining with Region Prototypical Network for Weakly Supervised Segmentation The code of: Cross-Image Region Mining with Region Proto

16 Nov 26, 2022

Weakly Supervised Posture Mining with Reverse Cross-entropy for Fine-grained Classification

Fine-grainedImageClassification Weakly Supervised Posture Mining with Reverse Cross-entropy for Fine-grained Classification We trained model here: lin

14 Oct 21, 2022

SSD: A Unified Framework for Self-Supervised Outlier Detection [ICLR 2021]

SSD: A Unified Framework for Self-Supervised Outlier Detection [ICLR 2021] Pdf: https://openreview.net/forum?id=v5gjXpmR8J Code for our ICLR 2021 pape

113 Nov 27, 2022

Yolo object detection - Yolo object detection with python

How to run download required files make build_image make download Docker versio

3 Jan 26, 2022

Official PyTorch implementation of "Proxy Synthesis: Learning with Synthetic Classes for Deep Metric Learning" (AAAI 2021)

Proxy Synthesis: Learning with Synthetic Classes for Deep Metric Learning Official PyTorch implementation of "Proxy Synthesis: Learning with Synthetic

30 Dec 6, 2022

Implementation of "Debiasing Item-to-Item Recommendations With Small Annotated Datasets" (RecSys '20)

Debiasing Item-to-Item Recommendations With Small Annotated Datasets This is the code for our RecSys '20 paper. Other materials can be found here: Ful

34 Aug 10, 2022

A PyTorch-based open-source framework that provides methods for improving the weakly annotated data and allows researchers to efficiently develop and compare their own methods.

Knodle (Knowledge-supervised Deep Learning Framework) - a new framework for weak supervision with neural networks. It provides a modularization for se

93 Nov 6, 2022

Annotated, understandable, and visually interpretable PyTorch implementations of: VAE, BIRVAE, NSGAN, MMGAN, WGAN, WGANGP, LSGAN, DRAGAN, BEGAN, RaGAN, InfoGAN, fGAN, FisherGAN

Overview PyTorch 0.4.1 | Python 3.6.5 Annotated implementations with comparative introductions for minimax, non-saturating, wasserstein, wasserstein g

471 Dec 16, 2022

Co-mining: Self-Supervised Learning for Sparsely Annotated Object Detection, AAAI 2021.

Related tags

Overview

Co-mining: Self-Supervised Learning for Sparsely Annotated Object Detection

Introduction

Get Started

Results

Citing Co-mining

You might also like...

This is the repository for the AAAI 21 paper [Contrastive and Generative Graph Convolutional Networks for Graph-based Semi-Supervised Learning].

Cross-Image Region Mining with Region Prototypical Network for Weakly Supervised Segmentation

Weakly Supervised Posture Mining with Reverse Cross-entropy for Fine-grained Classification

SSD: A Unified Framework for Self-Supervised Outlier Detection [ICLR 2021]

Yolo object detection - Yolo object detection with python

Official PyTorch implementation of "Proxy Synthesis: Learning with Synthetic Classes for Deep Metric Learning" (AAAI 2021)

Implementation of "Debiasing Item-to-Item Recommendations With Small Annotated Datasets" (RecSys '20)

A PyTorch-based open-source framework that provides methods for improving the weakly annotated data and allows researchers to efficiently develop and compare their own methods.

Annotated, understandable, and visually interpretable PyTorch implementations of: VAE, BIRVAE, NSGAN, MMGAN, WGAN, WGANGP, LSGAN, DRAGAN, BEGAN, RaGAN, InfoGAN, fGAN, FisherGAN

Comments

Working with two stage networks

Owner

MEGVII Research

Official PyTorch implementation of "Physics-aware Difference Graph Networks for Sparsely-Observed Dynamics".

The Self-Supervised Learner can be used to train a classifier with fewer labeled examples needed using self-supervised learning.

3D AffordanceNet is a 3D point cloud benchmark consisting of 23k shapes from 23 semantic object categories, annotated with 56k affordance annotations and covering 18 visual affordance categories.

Novel Instances Mining with Pseudo-Margin Evaluation for Few-Shot Object Detection

CCPD: a diverse and well-annotated dataset for license plate detection and recognition

In this project, we develop a face recognize platform based on MTCNN object-detection netcwork and FaceNet self-supervised network.

Unified Pre-training for Self-Supervised Learning and Supervised Learning for ASR

git《Self-Attention Attribution: Interpreting Information Interactions Inside Transformer》(AAAI 2021) GitHub:

Code for our NeurIPS 2021 paper Mining the Benefits of Two-stage and One-stage HOI Detection

[CVPR 2021] "The Lottery Tickets Hypothesis for Supervised and Self-supervised Pre-training in Computer Vision Models" Tianlong Chen, Jonathan Frankle, Shiyu Chang, Sijia Liu, Yang Zhang, Michael Carbin, Zhangyang Wang