• Image Recognition
• 2D Object Detection
• 3D Object Detection
• image retrieval
• metric learning, face recognition
• person/vehicle re-identification
• instance segmentation
• semantic segmentation
• tracking
• motion prediction for autonomous driving
• human object interaction detection
• pose estimation
• knowledge distillation
• domain adaptation

Spatially Adaptive Inference with Stochastic Feature Sampling and Interpolation

• paper: https://arxiv.org/abs/2003.08866
• summary:

MutualNet: Adaptive ConvNet via Mutual Learning from Network Width and Resolution

• paper: https://arxiv.org/abs/1909.12978
• summary:
• code: https://github.com/taoyang1122/MutualNet

Hybrid Models for Open Set Recognition

• paper: https://arxiv.org/abs/2003.12506
• summary:

Gradient Centralization: A New Optimization Technique for Deep Neural Networks

• paper: https://arxiv.org/abs/2004.01461
• summary:
• code: https://github.com/Yonghongwei/Gradient-Centralization

Multi-task Learning Increases Adversarial Robustness

• paper:
• summary:

Rethinking Bottleneck Structure for Efficient Mobile Network Design

• paper: https://arxiv.org/abs/2007.02269
• summary:

Negative Margin Matters: Understanding Margin in Few-shot Classification

• paper: https://arxiv.org/abs/2003.12060
• summary:
• code: https://github.com/bl0/negative-margin.few-shot

Dynamic Group Convolution for Accelerating Convolutional Neural Networks

• paper:
• summary:

PROFIT: A Novel Training Method for sub-4-bit MobileNet Models

• paper:
• summary:

URIE: Universal Image Enhancement for Visual Recognition in the Wild

• paper:
• summary:

DADA: Differentiable Automatic Data Augmentation

• paper: https://arxiv.org/abs/2003.03780
• summary:

AutoMix: Mixup Networks for Sample Interpolation via Cooperative Barycenter Learning

• paper:
• summary:

Momentum Batch Normalization for Deep Learning with Small Batch Size

• paper:
• summary:

Hard negatives examples are hard, but useful

• paper:
• summary:

Rethinking Few-Shot Image Classification: a Good Embedding Is All You Need?

• paper: https://arxiv.org/abs/2003.11539
• summary:

Topic-aware Multi-Label Classification

• paper:
• summary:

Resolution Switchable Networks for Runtime Efficient Image Classification

• paper:
• summary:

WeightNet: Revisiting the Design Space of Weight Networks

• paper:
• summary:

Suppressing Mislabeled Data via Grouping and Self-Attention

• paper:
• summary:

Attentive Normalization

• paper: https://arxiv.org/abs/1908.01259
• summary:

L2 Norm: A Generic Visualization Approach for Convolutional Neural Networks

• paper:
• summary:

FeatMatch: Feature-Based Augmentation for Semi-Supervised Learning

• paper:
• summary:

Fine-Grained Visual Classification via Progressive Multi-Granularity Training of Jigsaw Patches

• paper: https://arxiv.org/abs/2003.03836
• summary:

PSConv: Squeezing Feature Pyramid into One Compact Poly-Scale Convolutional Layer

• paper:
• summary:

Faster AutoAugment: Learning Augmentation Strategies using Backpropagation

• paper: https://arxiv.org/abs/1911.06987
• summary:

ATCluster: Self-Supervised Gaussian-Attention Network for Image Clustering

• paper:
• summary:

End-to-End Object Detection with Transformers

• paper: https://arxiv.org/abs/2005.12872
• summary:
• code: https://github.com/facebookresearch/detr

BorderDet: Border Feature for Dense Object Detection

• paper:
• summary:

Side-Aware Boundary Localization for More Precise Object Detection

• paper: https://arxiv.org/abs/1912.04260
• summary:
• code: https://github.com/open-mmlab/mmdetection

PIoU Loss: Towards Accurate Oriented Object Detection in Complex Environments

• paper:
• summary:

AABO: Adaptive Anchor Box Optimization for Object Detection via Bayesian Sub-sampling

• paper:
• summary:

A General Toolbox for Understanding Errors in Object Detection

• paper:
• summary:

Soft Anchor-Point Object Detection

• paper: https://arxiv.org/abs/1911.12448
• summary:

MimicDet: Bridging the Gap Between One-Stage and Two-Stage Object Detection

• paper:
• summary:

Disentangled Non-local Neural Networks

• paper: https://arxiv.org/abs/2006.06668
• summary:

Dynamic R-CNN: Towards High Quality Object Detection via Dynamic Training

• paper: https://arxiv.org/abs/2004.06002
• summary:

OS2D: One-Stage One-Shot Object Detection by Matching Anchor Features

• paper: https://arxiv.org/abs/2003.06800
• summary:
• code: https://github.com/aosokin/os2d

New Threats against Object Detector with Non-Local Block

• paper:
• summary:

Dense RepPoints: Representing Visual Objects with Dense Point Sets

• paper: https://arxiv.org/abs/1912.11473
• summary:
• code: https://github.com/justimyhxu/Dense-RepPoints

Large Batch Optimization for Object Detection: Training COCO in 12 Minutes

• paper:
• summary:

Dive Deeper Into Box for Object Detection

• paper:
• summary:

CenterNet Heatmap Propagation for Real-time Video Object Detection

• paper:
• summary:

Probabilistic Anchor Assignment with IoU Prediction for Object Detection

• paper:
• summary:

HoughNet: Integrating near and long-range evidence for bottom-up object detection

• paper: https://arxiv.org/abs/2007.02355
• summary:
• code: https://github.com/nerminsamet/houghnet

Domain Adaptive Object Detection via Asymmetric Tri-way Faster-RCNN

• paper: https://arxiv.org/abs/2007.01571
• summary:

Point-Set Anchors for Object Detection, Instance Segmentation and Pose Estimation

• paper: https://arxiv.org/abs/2007.02846
• summary:

RTM3D: Real-time Monocular 3D Detection from Object Keypoints for Autonomous Driving

• paper: https://arxiv.org/abs/2001.03343
• summary:

Finding Your (3D) Center: 3D Object Detection Using a Learned Loss

• paper: https://arxiv.org/abs/2004.02693
• summary:

Rotation-robust Intersection over Union for 3D Object Detection

• paper:
• summary:

Object as Hotspots: An Anchor-Free 3D Object Detection Approach via Firing of Hotspots

• paper: https://arxiv.org/abs/1912.12791
• summary:

Automated Data Augmentation Significantly Improves 3D Object Detection

• paper:
• summary:

Learning and aggregating deep local descriptors for instance-level recognition

• paper:
• summary:

Targeted Attack for Deep Hashing based Retrieval

• paper: https://arxiv.org/abs/2004.07955
• summary:

Online Invariance Selection for Local Feature Descriptors

• paper:
• summary:

ExchNet: A Unified Hashing Network for Large-Scale Fine-Grained Image Retrieval

• paper:
• summary:

S2DNet: Learning accurate correspondences for sparse-to-dense feature matching

• paper: https://arxiv.org/abs/2004.01673
• summary:

Unifying Deep Local and Global Features for Image Search

• paper: https://arxiv.org/abs/2001.05027
• summary:

SOLAR: Second-Order Loss and Attention for Image Retrieval

• paper: https://arxiv.org/abs/2001.08972
• summary:

Metric learning: cross-entropy vs. pairwise losses

• paper: https://arxiv.org/abs/2003.08983
• summary:

pherical Feature Transform for Deep Metric Learning

• paper:
• summary:

The Group Loss for Deep Metric Learning

• paper: http://arxiv.org/abs/1912.00385
• summary:

BroadFace: Looking at Tens of Thousands of People at Once for Face Recognition

• paper:
• summary:

Explainable Face Recognition

• paper:
• summary:

Joint Disentangling and Adaptation for Cross-Domain Person Re-Identification

• paper:
• summary:

Orientation-aware Vehicle Re-identification with Semantics-guided Part Attention Network

• paper:
• summary:

Identity-Guided Human Semantic Parsing Learning for Person Re-Identification

• paper:
• summary:

Faster Person Re-Identification

• paper:
• summary:

The Devil is in the Details: Self-Supervised Attention for Vehicle Re-Identification

• paper: https://arxiv.org/abs/2004.06271
• summary:

Generalizing Person Re-Identification by Camera-Aware Invariance Learning and Cross-Domain Mixup

• paper:
• summary:

Prediction, Recovery and Identification: Adaptive Low-Resolution Person Re-Identification

• paper:
• summary:

Multiple Expert Brainstorming for Domain Adaptive Person Re-identification

• paper: https://arxiv.org/abs/2007.01546
• summary:
• code: https://github.com/YunpengZhai/MEB-Net

Conditional Convolutions for Instance Segmentation

• paper: https://arxiv.org/abs/2003.05664
• summary:

SIP: Spatial Information Preservation for Fast Instance Segmentation

• paper:
• summary:

Learning with Noisy Class Labels for Instance Segmentation

• paper:
• summary:

Boundary-preserving Mask R-CNN

• paper:
• summary:

The Devil is in Classification: A Simple Framework for Long-tail Instance Segmentation

• paper:
• summary:

SOLO: Segmenting Objects by Locations

• paper: https://arxiv.org/abs/1912.04488
• summary:

LevelSet R-CNN: A Deep Variational Method for Instance Segmentation

• paper:
• summary:

PatchPerPix for Instance Segmentation

• paper: https://arxiv.org/abs/2001.07626
• summary:

Synthesize then Compare: Detecting Failures and Anomalies for Semantic Segmentation

• paper: https://arxiv.org/abs/2003.08440
• summary:

Semantic Flow for Fast and Accurate Scene Parsing

• paper: https://arxiv.org/abs/2002.10120
• summary:
• code: https://github.com/donnyyou/torchcv

Object-Contextual Representations for Semantic Segmentation

• paper: https://arxiv.org/abs/1909.11065
• summary:
• code: https://github.com/rosinality/ocr-pytorch

GINet: Graph Interaction Network for Scene Parsing

• paper:
• summary:

Blended Grammar Network for Human Parsing

• paper:
• summary:

EfficientFCN: Holistically-guided Decoding for Semantic Segmentation

• paper:
• summary:

Learning to Predict Context-adaptive Convolution for Semantic Segmentation

• paper: https://arxiv.org/abs/2004.08222
• summary:

SegFix: Model-Agnostic Boundary Refinement for Segmentation

• paper: https://arxiv.org/abs/2007.04269
• summary:
• code: https://github.com/openseg-group/openseg.pytorch

Segment as Points for Efficient Online Multi-Object Tracking and Segmentation

• paper: https://arxiv.org/abs/2007.01550
• summary:
• code: https://github.com/detectRecog/PointTrack

Chained-Tracker: Chaining Paired Attentive Regression Results for End-to-End Joint Multiple-Object Detection and Tracking

• paper:
• summary:

Tracking objects as points

• paper: https://arxiv.org/abs/2004.01177
• summary:
• code: https://github.com/xingyizhou/CenterTrack

TAO: A Large-scale Benchmark for Tracking Any Object

• paper: https://arxiv.org/abs/2005.10356
• summary:
• project page: http://taodataset.org/

Towards Real-time MOT: A Joint Solution for Detection and Appearance Embedding

• paper:
• summary:

Learning Feature Embeddings for Discriminant Model based Tracking

• paper:
• summary:

Learning Object-aware Anchor-free Networks for Real-time Object Tracking

• paper:
• summary:

PG-Net: Pixel to Global Matching Network for Visual Tracking Know Your Surroundings: Exploiting Scene Information for Object Tracking

• paper: https://arxiv.org/abs/2003.11014
• summary:

PiP: Planning-informed Trajectory Prediction for Autonomous Driving

• paper: https://arxiv.org/abs/2003.11476
• summary:

Detecting Human-Object Interactions with Action Co-occurrence Priors

• paper:
• summary:

UnionDet: Union-Level Detector Towards Real-Time Human-Object Interaction Detection

• paper:
• summary:

Visual Compositional Learning for Human Object Interaction Detection

• paper:
• summary:

Polysemy Deciphering Network for Human-Object Interaction Detection

• paper:
• summary:

Self6D: Self-Supervised Monocular 6D Object Pose Estimation

• paper: https://arxiv.org/abs/2004.06468
• summary:

End-to-End Estimation of Multi-Person 3D Poses from Multiple Cameras

• paper: https://arxiv.org/abs/2004.06239
• summary:

HMOR: Hierarchical Multi-person Ordinal Relations for Monocular Multi-Person 3D Pose Estimation

• paper:
• summary:

Learning Delicate Local Representations for Multi-Person Pose Estimation

• paper: http://arxiv.org/abs/2003.04030
• summary:

SMAP: Single-Shot Multi-Person Absolute 3D Pose Estimation

• paper:
• summary:

Adversarial Semantic Data Augmentation for Human Pose Estimation

• paper:
• summary:

Occlusion-Aware Siamese Network for Human Pose Estimation

• paper:
• summary:

emporal Keypoint Matching and Refinement Network for Pose Estimation and Tracking

• paper:
• summary:

Learning From Multiple Experts: Self-paced Knowledge Distillation for Long-tailed Classification

• paper: https://arxiv.org/abs/2001.01536
• summary:

Knowledge Transfer via Dense Cross-layer Mutual-distillation

• paper:
• summary:

Matching Guided Distillation

• paper:
• summary:

Feature Normalized Knowledge Distillation for Image Classification

• paper:
• summary:

Learning to Detect Open Classes for Universal Domain Adaptation

• paper:
• summary:

On the Effectiveness of Image Rotation for Open Set Domain Adapation

• paper:
• summary:

Self-Supervised CycleGAN for Object-Preserving Image-to-Image Domain Adaptation

• paper:
• summary:

Domain Adaptation through Task Distillation

• paper:
• summary: