Random Erasing Data Augmentation

Abstract

In this paper, we introduce Random Erasing, a new data augmentation method for training the convolutional neural network (CNN). In training, Random Erasing randomly selects a rectangle region in an image and erases its pixels with random values. In this process, training images with various levels of occlusion are generated, which reduces the risk of over-fitting and makes the model robust to occlusion. Random Erasing is parameter learning free, easy to implement, and can be integrated with most of the CNN-based recognition models. Albeit simple, Random Erasing is complementary to commonly used data augmentation techniques such as random cropping and flipping, and yields consistent improvement over strong baselines in image classification, object detection and person re-identification. Code is available at: https://github.com/zhunzhong07/Random-Erasing.

References in corpus (10)

• Fashion-MNIST: a Novel Image Dataset for Benchmarking Machine Learning Algorithms
• In Defense of the Triplet Loss for Person Re-Identification
• Wide Residual Networks
• Person Re-identification: Past, Present and Future
• Improving Person Re-identification by Attribute and Identity Learning
• Stochastic Pooling for Regularization of Deep Convolutional Neural Networks
• Pedestrian Alignment Network for Large-scale Person Re-identification
• Aggregated Residual Transformations for Deep Neural Networks
• Unlabeled Samples Generated by GAN Improve the Person Re-identification Baseline in vitro
• SVDNet for Pedestrian Retrieval

Cited by in corpus (201)

• YOLOv4: Optimal Speed and Accuracy of Object Detection
• AutoAugment: Learning Augmentation Policies from Data
• CutMix: Regularization Strategy to Train Strong Classifiers with Localizable Features
• AugMix: A Simple Data Processing Method to Improve Robustness and Uncertainty
• Data Augmentation using Random Image Cropping and Patching for Deep CNNs
• A Simple Semi-Supervised Learning Framework for Object Detection
• Patch-based Output Space Adversarial Learning for Joint Optic Disc and Cup Segmentation
• GridMask Data Augmentation
• MHSA-Net: Multi-Head Self-Attention Network for Occluded Person Re-Identification
• Data Augmentation for Skin Lesion Analysis
• Adaptive Exploration for Unsupervised Person Re-Identification
• A New Dataset, Poisson GAN and AquaNet for Underwater Object Grabbing
• Bag of Tricks and A Strong Baseline for Deep Person Re-identification
• Omni-Scale Feature Learning for Person Re-Identification
• Generative AI for Synthetic Data Across Multiple Medical Modalities: A Systematic Review of Recent Developments and Challenges
• NSML: Meet the MLaaS platform with a real-world case study
• FMix: Enhancing Mixed Sample Data Augmentation
• Multi-task Learning with Coarse Priors for Robust Part-aware Person Re-identification
• CSWin Transformer: A General Vision Transformer Backbone with Cross-Shaped Windows
• Pose-Aware Instance Segmentation Framework from Cone Beam CT Images for Tooth Segmentation
• Hierarchy-based Image Embeddings for Semantic Image Retrieval
• Improving Robustness Without Sacrificing Accuracy with Patch Gaussian Augmentation
• Incomplete Descriptor Mining with Elastic Loss for Person Re-Identification
• Assuring the Machine Learning Lifecycle: Desiderata, Methods, and Challenges
• Parameter-Free Spatial Attention Network for Person Re-Identification
• Dual Attention Matching Network for Context-Aware Feature Sequence based Person Re-Identification
• Lightweight Multi-Branch Network for Person Re-Identification
• Privacy-preserving Machine Learning through Data Obfuscation
• ASAP: Architecture Search, Anneal and Prune
• ABD-Net: Attentive but Diverse Person Re-Identification
• TripletTrack: 3D Object Tracking using Triplet Embeddings and LSTM
• Data Augmentation for Object Detection via Progressive and Selective Instance-Switching
• Invariance Matters: Exemplar Memory for Domain Adaptive Person Re-identification
• Image-Image Domain Adaptation with Preserved Self-Similarity and Domain-Dissimilarity for Person Re-identification
• Affinity and Diversity: Quantifying Mechanisms of Data Augmentation
• How Robust is 3D Human Pose Estimation to Occlusion?
• Inspect, Understand, Overcome: A Survey of Practical Methods for AI Safety
• Learning Generalisable Omni-Scale Representations for Person Re-Identification
• Self-similarity Grouping: A Simple Unsupervised Cross Domain Adaptation Approach for Person Re-identification
• VideoMix: Rethinking Data Augmentation for Video Classification
• An Ensemble of Simple Convolutional Neural Network Models for MNIST Digit Recognition
• A Sequence Matching Network for Polyphonic Sound Event Localization and Detection
• Lung Segmentation from Chest X-rays using Variational Data Imputation
• High-Order Information Matters: Learning Relation and Topology for Occluded Person Re-Identification
• Mixed High-Order Attention Network for Person Re-Identification
• Relation-Aware Global Attention for Person Re-identification
• Synthetic Occlusion Augmentation with Volumetric Heatmaps for the 2018 ECCV PoseTrack Challenge on 3D Human Pose Estimation
• Domain Adaptive Person Re-Identification via Coupling Optimization
• Identity-Guided Human Semantic Parsing for Person Re-Identification
• Learning to Adapt Invariance in Memory for Person Re-identification
• An Empirical Evaluation on Robustness and Uncertainty of Regularization Methods
• Camera Style Adaptation for Person Re-identification
• 'Skimming-Perusal' Tracking: A Framework for Real-Time and Robust Long-term Tracking
• Hierarchical Clustering with Hard-batch Triplet Loss for Person Re-identification
• ES-Net: Erasing Salient Parts to Learn More in Re-Identification
• Vehicle Attribute Recognition by Appearance: Computer Vision Methods for Vehicle Type, Make and Model Classification
• The Herbarium Challenge 2019 Dataset
• Looking GLAMORous: Vehicle Re-Id in Heterogeneous Cameras Networks with Global and Local Attention
• Deep Group-shuffling Random Walk for Person Re-identification
• Batch DropBlock Network for Person Re-identification and Beyond
• Cross-modality Person re-identification with Shared-Specific Feature Transfer
• Towards Good Practices on Building Effective CNN Baseline Model for Person Re-identification
• Towards Robust Learning-Based Pose Estimation of Noncooperative Spacecraft
• Mixing Real and Synthetic Data to Enhance Neural Network Training -- A Review of Current Approaches
• Regularizing Deep Networks with Semantic Data Augmentation
• XNAS: Neural Architecture Search with Expert Advice
• GraphCrop: Subgraph Cropping for Graph Classification
• Lightweight Residual Densely Connected Convolutional Neural Network
• Randomness In Neural Network Training: Characterizing The Impact of Tooling
• Part-Guided Attention Learning for Vehicle Instance Retrieval
• CA3Net: Contextual-Attentional Attribute-Appearance Network for Person Re-Identification
• Progressive DARTS: Bridging the Optimization Gap for NAS in the Wild
• Beyond Triplet Loss: Person Re-identification with Fine-grained Difference-aware Pairwise Loss
• Beyond Human Parts: Dual Part-Aligned Representations for Person Re-Identification
• Densely Semantically Aligned Person Re-Identification
• Person Re-identification with Deep Similarity-Guided Graph Neural Network
• Learning Dynamic Alignment via Meta-filter for Few-shot Learning
• Locality Aware Appearance Metric for Multi-Target Multi-Camera Tracking
• Temporal Complementary Learning for Video Person Re-Identification
• Exchangeable deep neural networks for set-to-set matching and learning
• Joint Noise-Tolerant Learning and Meta Camera Shift Adaptation for Unsupervised Person Re-Identification
• GiT: Graph Interactive Transformer for Vehicle Re-identification
• A heterogeneous branch and multi-level classification network for person re-identification
• KeepAugment: A Simple Information-Preserving Data Augmentation Approach
• Modeling Visual Context is Key to Augmenting Object Detection Datasets
• Multi-Granularity Reference-Aided Attentive Feature Aggregation for Video-based Person Re-identification
• View Confusion Feature Learning for Person Re-identification
• Occluded Person Re-identification
• Learning data augmentation policies using augmented random search
• Skin lesion classification with ensemble of squeeze-and-excitation networks and semi-supervised learning
• Monocular UAV Localisation with Deep Learning and Uncertainty Propagation
• Data Augmentation with Manifold Exploring Geometric Transformations for Increased Performance and Robustness
• Spectral Feature Transformation for Person Re-identification
• Dual-Refinement: Joint Label and Feature Refinement for Unsupervised Domain Adaptive Person Re-Identification
• Spatial-Temporal Person Re-identification
• FenceMask: A Data Augmentation Approach for Pre-extracted Image Features
• Semantics-Aligned Representation Learning for Person Re-identification
• Adaptively Connected Neural Networks
• Query Attack via Opposite-Direction Feature:Towards Robust Image Retrieval
• One-Shot GAN Generated Fake Face Detection
• Segmentation-driven 6D Object Pose Estimation
• Black Re-ID: A Head-shoulder Descriptor for the Challenging Problem of Person Re-Identification
• SnapMix: Semantically Proportional Mixing for Augmenting Fine-grained Data
• Adaptive Label Smoothing
• A Similarity Inference Metric for RGB-Infrared Cross-Modality Person Re-identification
• Circumventing Outliers of AutoAugment with Knowledge Distillation
• Robustness to Augmentations as a Generalization metric
• An Evaluation of Deep CNN Baselines for Scene-Independent Person Re-Identification
• A Contour Stochastic Gradient Langevin Dynamics Algorithm for Simulations of Multi-modal Distributions
• On the Importance of Visual Context for Data Augmentation in Scene Understanding
• When Human Pose Estimation Meets Robustness: Adversarial Algorithms and Benchmarks
• Factorized Distillation: Training Holistic Person Re-identification Model by Distilling an Ensemble of Partial ReID Models
• Grafted network for person re-identification
• Neural Networks and Polynomial Regression. Demystifying the Overparametrization Phenomena
• Fast and Accurate Person Re-Identification with RMNet
• Parallel Grid Pooling for Data Augmentation
• Weakly Supervised Person Re-ID: Differentiable Graphical Learning and A New Benchmark
• Unity Style Transfer for Person Re-Identification
• VINNAS: Variational Inference-based Neural Network Architecture Search
• End-to-End Deep Kronecker-Product Matching for Person Re-identification
• Consistent Sparse Deep Learning: Theory and Computation
• Dual Gaussian-based Variational Subspace Disentanglement for Visible-Infrared Person Re-Identification
• Can We Read Speech Beyond the Lips? Rethinking RoI Selection for Deep Visual Speech Recognition
• Meta-Learning Initializations for Image Segmentation
• Adaptive L2 Regularization in Person Re-Identification
• Every Node Counts: Self-Ensembling Graph Convolutional Networks for Semi-Supervised Learning
• Improving Object Detection with Inverted Attention
• Multi-Scale Body-Part Mask Guided Attention for Person Re-identification
• GreyReID: A Two-stream Deep Framework with RGB-grey Information for Person Re-identification
• Long-Term Cloth-Changing Person Re-identification
• Uncertainty-Aware Multi-Shot Knowledge Distillation for Image-Based Object Re-Identification
• Backbone Can Not be Trained at Once: Rolling Back to Pre-trained Network for Person Re-Identification
• A Preliminary Study on Data Augmentation of Deep Learning for Image Classification
• BiCnet-TKS: Learning Efficient Spatial-Temporal Representation for Video Person Re-Identification
• End-to-End Text Classification via Image-based Embedding using Character-level Networks
• Does Data Augmentation Benefit from Split BatchNorms
• A Comprehensive Study of Data Augmentation Strategies for Prostate Cancer Detection in Diffusion-weighted MRI using Convolutional Neural Networks
• Unsupervised Eyeglasses Removal in the Wild
• Style transfer-based image synthesis as an efficient regularization technique in deep learning
• Single-Stage 6D Object Pose Estimation
• Angular Triplet Loss-based Camera Network for ReID
• Relation Network for Person Re-identification
• Faster Person Re-Identification
• Improved Res2Net model for Person re-identification
• IAUnet: Global Context-Aware Feature Learning for Person Re-Identification
• Branch-Cooperative OSNet for Person Re-Identification
• In Defense of the Classification Loss for Person Re-Identification
• Fine-Grained Re-Identification
• HAVANA: Hierarchical and Variation-Normalized Autoencoder for Person Re-identification
• Augmented Cyclic Consistency Regularization for Unpaired Image-to-Image Translation
• Why Mixup Improves the Model Performance
• SSAH: Semi-supervised Adversarial Deep Hashing with Self-paced Hard Sample Generation
• Scalable and Practical Natural Gradient for Large-Scale Deep Learning
• Cluster-level Feature Alignment for Person Re-identification
• Intra-clip Aggregation for Video Person Re-identification
• Homocentric Hypersphere Feature Embedding for Person Re-identification
• Fingerprint Feature Extraction by Combining Texture, Minutiae, and Frequency Spectrum Using Multi-Task CNN
• Data augmentation with Mobius transformations
• Semi-supervised learning for medical image classification using imbalanced training data
• Exploring Vision Transformers for Fine-grained Classification
• Hierarchical and Efficient Learning for Person Re-Identification
• Spatial-Aware GAN for Unsupervised Person Re-identification
• Progressive Learning Algorithm for Efficient Person Re-Identification
• Image-based Vehicle Re-identification Model with Adaptive Attention Modules and Metadata Re-ranking
• End-to-End Training of CNN Ensembles for Person Re-Identification
• Robust, Extensible, and Fast: Teamed Classifiers for Vehicle Tracking and Vehicle Re-ID in Multi-Camera Networks
• Searching for A Robust Neural Architecture in Four GPU Hours
• Sparse Label Smoothing Regularization for Person Re-Identification
• Cross-Correlated Attention Networks for Person Re-Identification
• Augmented Data as an Auxiliary Plug-in Towards Categorization of Crowdsourced Heritage Data
• ROI Regularization for Semi-supervised and Supervised Learning
• Learning Augmentation Distributions using Transformed Risk Minimization
• Medical Image Segmentation with Limited Supervision: A Review of Deep Network Models
• Fast Mesh Data Augmentation via Chebyshev Polynomial of Spectral filtering
• Learning Deep Representations by Mutual Information for Person Re-identification
• Hybrid Attention for Automatic Segmentation of Whole Fetal Head in Prenatal Ultrasound Volumes
• Progressive Multi-stage Feature Mix for Person Re-Identification
• Pose-Guided Feature Learning with Knowledge Distillation for Occluded Person Re-Identification
• Mixing between the Cross Entropy and the Expectation Loss Terms
• Adaptive Deep Metric Embeddings for Person Re-Identification under Occlusions
• Push for Center Learning via Orthogonalization and Subspace Masking for Person Re-Identification
• Automated Segmentation of Hip and Thigh Muscles in Metal Artifact-Contaminated CT using Convolutional Neural Network-Enhanced Normalized Metal Artifact Reduction
• Water Supply Prediction Based on Initialized Attention Residual Network
• Learning Incremental Triplet Margin for Person Re-identification
• What Else Can Fool Deep Learning? Addressing Color Constancy Errors on Deep Neural Network Performance
• Differentiable Supervector Extraction for Encoding Speaker and Phrase Information in Text Dependent Speaker Verification
• ShuffleBlock: Shuffle to Regularize Deep Convolutional Neural Networks
• Pose Invariant Person Re-Identification using Robust Pose-transformation GAN
• A Generalization Theory based on Independent and Task-Identically Distributed Assumption
• Cross-Resolution Person Re-identification with Deep Antithetical Learning
• Learning Inward Scaled Hypersphere Embedding: Exploring Projections in Higher Dimensions
• Progressive Bilateral-Context Driven Model for Post-Processing Person Re-Identification
• Survey on Deep Learning-based Kuzushiji Recognition
• Accurate Scene Text Detection through Border Semantics Awareness and Bootstrapping
• Channel Recurrent Attention Networks for Video Pedestrian Retrieval
• Cooperative Bi-path Metric for Few-shot Learning
• Physical world assistive signals for deep neural network classifiers -- neither defense nor attack
• FlipReID: Closing the Gap between Training and Inference in Person Re-Identification
• Unified Multifaceted Feature Learning for Person Re-Identification
• VMRFANet:View-Specific Multi-Receptive Field Attention Network for Person Re-identification
• Towards Solving the DeepFake Problem : An Analysis on Improving DeepFake Detection using Dynamic Face Augmentation