/TopPaper

Classic papers for beginners, and impact scope for authors.

MIT LicenseMIT

TopPaper

Classic Papers for Beginners and Impact Scope for Authors.

There have been billions of academic papers around the world. However, maybe only 0.0...01% among them are valuable or are worth reading. Since our limited life has never been forever, TopPaper provide a Top Academic Paper Chart for beginners and reseachers to take one step faster.

Welcome to contribute more subject or valuable (at least you think) papers. Please feel free to pull requests or open an issue.



0. Traditional Methods

Abbreviation Paper Cited by Journal Year 1st Author 1st Affiliation
SIFT Object Recognition from Local Scale-Invariant Features 20 K ICCV 1999 David G. Lowe University of British Columbia
HOG Histograms of Oriented Gradients for Human Detection 35 K CVPR 2005 Navneet Dalal inrialpes
SURF SURF: Speeded Up Robust Features 18 K ECCV 2006 Herbert Bay ETH Zurich
......

1. CNN [Convolutional Neural Network]

1.1 Image Classification

1.1.1 Architecture

Abbreviation Paper Cited By Journal Year 1st Author 1st Affiliation
LeNet Backpropagation applied to handwritten zip code recognition 8.3 K Neural Computation 1989 Yann Lecun AT&T Bell Laboratories
LeNet Gradient-based learning applied to document recognition 35 K Proceedings of the IEEE 1998 Yann Lecun AT&T Research Laboratories
ImageNet ImageNet: A large-scale hierarchical image database 26 K CVPR 2009 Jia Dengn Princeton University
AlexNet ImageNet Classification with Deep Convolutional Neural Networks 79 K NIPS 2012 Alex Krizhevsky University of Toronto
ZFNet Visualizing and Understanding Convolutional Networks 11 K ECCV 2014 Matthew D Zeiler New York University
VGGNet Very Deep Convolutional Networks for Large-Scale Image Recognition 55 K ICLR 2015 Karen Simonyan Oxford
GoogLeNet Going Deeper with Convolutions 29 K CVPR 2015 Christian Szegedy Google
GoogLeNet_v2_v3 Rethinking the Inception Architecture for Computer Vision 12 K CVPR 2016 Christian Szegedy Google
ResNet Deep Residual Learning for Image Recognition 74 K CVPR 2016 Kaiming He MSRA
DenseNet Densely Connected Convolutional Networks 15 K CVPR 2017 Gao Huang Cornell University
ResNeXt Aggregated Residual Transformations for Deep Neural Networks 3.9 K CVPR 2017 Saining Xie UC San Diego
MobileNet MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applications 7.7 K arXiv 2017 Andrew G. Howard Google
SENet Squeeze-and-Excitation Networks 6.3 K CVPR 2018 Jie Hu Momenta
MobileNet_v2 MobileNetV2: Inverted Residuals and Linear Bottlenecks 4.4 K CVPR 2018 Mark Sandler Google
ShuffleNet ShuffleNet: An Extremely Efficient Convolutional Neural Network for Mobile Devices 2.3 K CVPR 2018 Xiangyu Zhang Megvii
ShuffleNet V2 ShuffleNet V2: Practical Guidelines for Efficient CNN Architecture Design 1.3 K ECCV 2018 Ningning Ma Megvii
MobileNet_v3 Searching for MobileNetV3 0.6 K ICCV 2019 Andrew Howard Google
EfficientNet EfficientNet: Rethinking Model Scaling for Convolutional Neural Networks 1.9 K ICML 2019 Mingxing Tan Google
GhostNet GhostNet: More Features from Cheap Operations 0.1 K CVPR 2020 Kai Han Huawei Noah
AdderNet AdderNet: Do We Really Need Multiplications in Deep Learning? 33 CVPR 2020 Hanting Chen Huawei Noah
Res2Net Res2Net: A New Multi-scale Backbone Architecture 0.2 K TPAMI 2021 Shang-Hua Gao Nankai University

1.1.2 Dataset, Augmentation, Trick

Abbreviation Paper Cited By Journal Year 1st Author 1st Affiliation
BN Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift 26 K ICML 2015 Sergey Ioffe Google
Xavier Init Understanding the difficulty of training deep feedforward neural networks 12 K AISTATS 2010 Xavier Universite de Montreal
Kaiming Init Delving Deep into Rectifiers: Surpassing Human-Level Performance on ImageNet Classification 11 K ICCV 2015 Kaiming He MSRA
LN Layer Normalization 2.9 K NIPS 2016 Jimmy Lei Ba University of Toronto
GN Group Normalization 1.1 K ECCV 2018 Yuxin Wu FAIR
- Bag of Tricks for Image Classification with Convolutional Neural Networks 361 CVPR 2019 Tong He Amazon
- Fixing the train-test resolution discrepancy 122 NeurIPS 2019 Hugo Touvron FAIR
Auto-Augment AutoAugment: Learning Augmentation Policies from Data 487 CVPR 2019 Ekin D. Cubuk Google
- Fixing the train-test resolution discrepancy: FixEfficientNet 53 Arxiv 2020 Hugo Touvron FAIR

1.2 Object Detection

Abbreviation Paper Cited By Journal Year 1st Author 1st Affiliation
RCNN Rich feature hierarchies for accurate object detection and semantic segmentation 17 K CVPR 2014 Ross Girshick Berkeley
Fast RCNN Fast R-CNN 14 K ICCV 2015 Ross Girshick Microsoft Research
Faster RCNN Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks 20 K NIPS 2015 Shaoqing Ren USTC, MSRA
SSD SSD: Single Shot MultiBox Detector 13 K ECCV 2016 Wei Liu UNC
YOLO You Only Look Once: Unified, Real-Time Object Detection 15 K CVPR 2016 Joseph Redmon University of Washington
Mask RCNN Mask R-CNN 10 K ICCV 2017 Kaiming He FAIR
DSSD DSSD : Deconvolutional Single Shot Detector 1.0 K CVPR 2017 Cheng-Yang Fu UNC
YOLO9000 YOLO9000: Better, Faster, Stronger. 7.7 K CVPR 2017 Joseph Redmon University of Washington
FPN Feature Pyramid Networks for Object Detection 6.7 K CVPR 2017 Tsung-Yi Lin FAIR
Focal Loss Focal Loss for Dense Object Detection 6.7 K ICCV 2017 Tsung-Yi Lin FAIR
Deformable Conv Deformable Convolutional Networks 1.6 K ICCV 2017 Jifeng Dai MSRA
YOLO V3 Yolov3: An incremental improvement 6.9 K CVPR 2018 Joseph Redmon University of Washington
ATSS Bridging the Gap Between Anchor-based and Anchor-free Detection via Adaptive Training Sample Selection 0.1 K CVPR 2020 Shifeng Zhang CASIA
EfficientDet EfficientDet: Scalable and Efficient Object Detection 0.3 K CVPR 2020 Mingxing Tan Google

1.3 Object Segmentation

Abbreviation Paper Cited By Journal Year 1st Author 1st Affiliation
FCN Fully Convolutional Networks for Semantic Segmentation 22 K CVPR 2015 Jonathan Long UC Berkeley
DeepLab DeepLab: Semantic Image Segmentation with Deep Convolutional Nets, Atrous Convolution, and Fully Connected CRFs 7.4 K ICLR 2015 Liang-Chieh Chen Google
Unet U-Net: Convolutional Networks for Biomedical Image Segmentation 24 K MICCAI 2015 Olaf Ronneberger University of Freiburg
- Learning to Segment Object Candidates 0.6 K NIPS 2015 Pedro O. Pinheiro FAIR
Dilated Conv Multi-Scale Context Aggregation by Dilated Convolutions 4.5 K ICLR 2016 Fisher Y Princeton University
- Large Kernel Matters -- Improve Semantic Segmentation by Global Convolutional Network 0.7 K CVPR 2017 Chao Peng Tsinghua
RefineNet RefineNet: Multi-Path Refinement Networks for High-Resolution Semantic Segmentation 1.6 K CVPR 2017 Guosheng Lin The University of Adelaide

1.4 Re_ID [Person Re-Identification]

1.5 OCR [Optical Character Recognition]

Abbreviation Paper Cited by Journal Year 1st Author 1st Affiliation
CTC Connectionist temporal classifaction: labelling unsegmented sequence data with recurrent neural network 2.9 K ICML 2006 Alex Graves IDSIA

1.6 Face Recognition

Abbreviation Paper Cited by Journal Year 1st Author 1st Affiliation
DeepFace DeepFace: Closing the Gap to Human-Level Performance in Face Verification 5.3 K CVPR 2014 Yaniv Taigman FAIR
DeepID v1 Deep Learning Face Representation from Predicting 10,000 Classes 1.8 K CVPR 2014 Yi Sun CUHK
DeepID v2 Deep Learning Face Representation by Joint Identification-Verification 1.9 K NIPS 2014 Yi Sun CUHK
FaceNet FaceNet: A Unified Embedding for Face Recognition and Clustering 7.4 K CVPR 2015 Florian Schrof Google
Center Loss A Discriminative Feature Learning Approach for Deep Face Recognition 2.1 K ECCV 2016 Yandong Wen CMU
ArcFace ArcFace: Additive Angular Margin Loss for Deep Face Recognition 1.3 K CVPR 2017 Jiankang Deng Imperial College London
SphereFace SphereFace: Deep Hypersphere Embedding for Face Recognition 1.3 K CVPR 2017 Weiyang Liu Georgia Institute of Technology
CosFace CosFace: Large Margin Cosine Loss for Deep Face Recognition 0.8 K CVPR 2018 Hao Wang Tecent
AM-Softmax Loss Additive Margin Softmax for Face Verification 0.5 K Signal Processing Letters 2018 Feng Wang UESTC

1.7 NAS [Neural Architecture Search]

Abbreviation Paper Cited By Journal Year 1st Author 1st Affiliation
Darts DARTS: Differentiable Architecture Search 1.3 K ICLR 2019 Hanxiao Liu CMU
- Neural Architecture Search with Reinforcement Learning 2.5 K ICLR 2017 Barret Zoph Google
- Efficient Neural Architecture Search via Parameter Sharing 1.2 K ICML 2018 Hieu Pham Google
- SNAS: Stochastic Neural Architecture Search 0.3 K ICLR 2019 Sirui Xie SenseTime
PC-Darts PC-DARTS: Partial Channel Connections for Memory-Efficient Architecture Search 159 ICLR 2020 Yuhui Xu Huawei

1.8 Image Super_Resolution

Abbreviation Paper Cited By Journal Year 1st Author 1st Affiliation
SRCNN Image Super-Resolution Using Deep Convolutional Networks 4.1 K ECCV 2014 Chao Dong CUHK
ESPCN Real-Time Single Image and Video Super-Resolution Using an Efficient Sub-Pixel Convolutional Neural Network 2.4 K CVPR 2016 Wenzhe Shi Twitter
FSRCNN Accelerating the Super-Resolution Convolutional Neural Network 1.3 K ECCV 2016 Chao Dong CUHK
VDSR Accurate Image Super-Resolution Using Very Deep Convolutional Networks 3.5 K CVPR 2016 Jiwon Kim Seoul National University
DRCN Deeply-Recursive Convolutional Network for Image Super-Resolution 1.4 K CVPR 2016 Jiwon Kim Seoul National University
EDSR Enhanced Deep Residual Networks for Single Image Super-Resolution 2.0 K CVPRW 2017 Bee Lim Seoul National University
DRRN Image Super-Resolution via Deep Recursive Residual Network 1.0 K CVPR 2017 Ying Tai NJUST
SRDenseNet Image Super-Resolution Using Dense Skip Connections 0.5 K ICCV 2017 Tong Tong Imperial Vision
SRGAN Photo-Realistic Single Image Super-Resolution Using a Generative Adversarial Network 5.3 K CVPR 2017 Christian Ledig Twitter
LapSRN Deep Laplacian Pyramid Networks for Fast and Accurate Super-Resolution 1.1 K CVPR 2017 Wei-Sheng Lai 1University of California
RDN Residual Dense Network for Image Super-Resolution 1.1 K CVPR 2018 Yulun Zhang Northeastern University
DBPN Deep Back-Projection Networks For Super-Resolution 0.6 K CVPR 2018 Muhammad Haris Toyota Technological Institute
RCAN Image Super-Resolution Using Very Deep Residual Channel Attention Networks 1.0 K ECCV 2018 Yulun Zhang Northeastern University

1.9 Image Denoising

Abbreviation Paper Cited By Journal Year 1st Author 1st Affiliation
CBDNet Toward Convolutional Blind Denoising of Real Photographs 0.2 K CVPR 2019 Shi Guo HIT
- Learning Deep CNN Denoiser Prior for Image Restoration 0.8 K CVPR 2017 Kai Zhang HIT
CnDNN Beyond a Gaussian Denoiser: Residual Learning of Deep CNN for Image Denoising 2.9 K TIP 2017 Kai Zhang HIT
FFDNet FFDNet: Toward a fast and flexible solution for CNN based image denoising 0.6 K TIP 2018 Kai Zhang HIT
SRMD Learning a Single Convolutional Super-Resolution Network for Multiple Degradations 0.3 K CVPR 2018 Kai Zhang HIT
RIDNet Real Image Denoising with Feature Attention] 87 ICCV 2019 Saeed Anwar CSIRO
CycleISP CycleISP: Real Image Restoration via Improved Data Synthesis 28 CVPR 2020 Syed Waqas Zamir UAE
AINDNet Transfer Learning from Synthetic to Real-Noise Denoising with Adaptive Instance Normalization 14 CVPR 2020 Yoonsik Kim Seoul National University

1.10 Model Compression, Pruning, Quantization, Knowledge Distillation

Abbreviation Paper Cited By Journal Year 1st Author 1st Affiliation
KD Distilling the Knowledge in a Neural Network 5.8 K NIPS-W 2014 Geoffrey Hinton Google
DeepCompression Deep Compression: Compressing Deep Neural Networks with Pruning, Trained Quantization and Huffman Coding 4.9K ICLR 2016 Song Han Stanford
Fixed Point Quant Fixed point quantization of deep convolutional networks 0.5 K ICLR-W 2016 Darryl D. Lin Qualcomm
DoReFa DoReFa-Net: Training Low Bitwidth Convolutional Neural Networks with Low Bitwidth Gradients 1.1 K CVPR 2016 Shuchang Zhou Megvii
Fake Quant Quantization and Training of Neural Networks for Efficient Integer-Arithmetic-Only Inference 0.8 K CVPR 2018 Benoit Jacob Google
Once for all Once-for-All: Train One Network and Specialize it for Efficient Deployment 0.1 K ICLR 2020 Han Cai MIT

2. Transformer in Vision

Abbreviation Paper Cited by Journal Year 1st Author 1st Affiliation
Image Transformer Image Transformer 337 ICML 2018 Niki Parmar Google
- Attention Augmented Convolutional Networks 191 ICCV 2019 Irwan Bello Google
DETR End-to-End Object Detection with Transformers 252 ECCV 2020 Nicolas Carion Facebook AI
Deit Training data-efficient image transformers & distillation through attention 57 arXiv 2020 Hugo Touvron FAIR
i-GPT Generative Pretraining from Pixels 38 ICML 2020 Mark Chen OpenAI
Deformable DETR Deformable DETR: Deformable Transformers for End-to-End Object Detection 12 ICLR 2021 Xizhou Zhu SenseTime
- Training data-efficient image transformers & distillation through attention 57 Arxiv 2020 Hugo Touvron FAIR
ViT An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale 175 ICLR 2021 Alexey Dosovitskiy Google
IPT Pre-Trained Image Processing Transformer 16 CVPR 2021 Hanting Chen Huawei Noah
- A Survey on Visual Transformer 12 Arxiv 2021 Kai Han Huawei Noah
TNT Transformer in Transformer 8 Arxiv 2021 Kai Han Huawei Noah
......

3. Transformer and Self-Attention in NLP

Abbreviation Paper Cited by Journal Year 1st Author 1st Affiliation
Transformer Attention Is All You Need 19 K NIPS 2017 Ashish Vaswani Google
- Self-Attention with Relative Position Representations 0.5 K NAACL 2018 Peter Shaw Google
Bert BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding 17 K NAACL 2019 Jacob Devlin Google

4. Others

......

Acknowledgement

Thanks for the materias and help from Aidong Men, Bo Yang, Zhuqing Jiang, Qishuo Lu, Zhengxin Zeng, Jia'nan Han, Pengliang Tang, Yiyun Zhao, Xian Zhang ......