Data Augmentation by Pairing Samples for Images Classification

Abstract

Data augmentation is a widely used technique in many machine learning tasks, such as image classification, to virtually enlarge the training dataset size and avoid overfitting. Traditional data augmentation techniques for image classification tasks create new samples from the original training data by, for example, flipping, distorting, adding a small amount of noise to, or cropping a patch from an original image. In this paper, we introduce a simple but surprisingly effective data augmentation technique for image classification tasks. With our technique, named SamplePairing, we synthesize a new sample from one image by overlaying another image randomly chosen from the training data (i.e., taking an average of two images for each pixel). By using two images randomly selected from the training set, we can generate $N^2$ new samples from $N$ training samples. This simple data augmentation technique significantly improved classification accuracy for all the tested datasets; for example, the top-1 error rate was reduced from 33.5% to 29.0% for the ILSVRC 2012 dataset with GoogLeNet and from 8.22% to 6.93% in the CIFAR-10 dataset. We also show that our SamplePairing technique largely improved accuracy when the number of samples in the training set was very small. Therefore, our technique is more valuable for tasks with a limited amount of training data, such as medical imaging tasks.

References in corpus (6)

• Adam: A Method for Stochastic Optimization
• SMOTE: Synthetic Minority Over-sampling Technique
• Improving neural networks by preventing co-adaptation of feature detectors
• mixup: Beyond Empirical Risk Minimization
• Learning from Between-class Examples for Deep Sound Recognition
• Between-class Learning for Image Classification

Cited by in corpus (60)

• BACH: Grand Challenge on Breast Cancer Histology Images
• AutoAugment: Learning Augmentation Policies from Data
• A Comprehensive Survey of Image Augmentation Techniques for Deep Learning
• Efficient Deep Learning: A Survey on Making Deep Learning Models Smaller, Faster, and Better
• A Survey on Data Augmentation for Text Classification
• EmergencyNet: Efficient Aerial Image Classification for Drone-Based Emergency Monitoring Using Atrous Convolutional Feature Fusion
• Recent Advances in Object Detection in the Age of Deep Convolutional Neural Networks
• A physics-informed operator regression framework for extracting data-driven continuum models
• FMix: Enhancing Mixed Sample Data Augmentation
• Review of data analysis in vision inspection of power lines with an in-depth discussion of deep learning technology
• An overview of mixing augmentation methods and augmentation strategies
• Deep Learning Methods for Solving Linear Inverse Problems: Research Directions and Paradigms
• Data Augmentation for Object Detection via Progressive and Selective Instance-Switching
• An End-to-End Breast Tumour Classification Model Using Context-Based Patch Modelling- A BiLSTM Approach for Image Classification
• Affinity and Diversity: Quantifying Mechanisms of Data Augmentation
• Multi-Sample Dropout for Accelerated Training and Better Generalization
• Inspect, Understand, Overcome: A Survey of Practical Methods for AI Safety
• Mixup Inference: Better Exploiting Mixup to Defend Adversarial Attacks
• DADA: Differentiable Automatic Data Augmentation
• Transductive Episodic-Wise Adaptive Metric for Few-Shot Learning
• Improving Auto-Augment via Augmentation-Wise Weight Sharing
• Mixing Real and Synthetic Data to Enhance Neural Network Training -- A Review of Current Approaches
• Sill-Net: Feature Augmentation with Separated Illumination Representation
• Deep Learning for Classification and Severity Estimation of Coffee Leaf Biotic Stress
• Data Augmentation for Deep Learning-based Radio Modulation Classification
• On the Generalization Effects of Linear Transformations in Data Augmentation
• Influence-guided Data Augmentation for Neural Tensor Completion
• Learning to Recombine and Resample Data for Compositional Generalization
• Neural Semi-supervised Learning for Text Classification Under Large-Scale Pretraining
• Image Obfuscation for Privacy-Preserving Machine Learning
• One-Shot GAN Generated Fake Face Detection
• Data Interpolating Prediction: Alternative Interpretation of Mixup
• Data Augmentation for Object Detection via Differentiable Neural Rendering
• SnapMix: Semantically Proportional Mixing for Augmenting Fine-grained Data
• Improved Adversarial Robustness via Logit Regularization Methods
• Human-Expert-Level Brain Tumor Detection Using Deep Learning with Data Distillation and Augmentation
• Neural Networks Are More Productive Teachers Than Human Raters: Active Mixup for Data-Efficient Knowledge Distillation from a Blackbox Model
• PK-GCN: Prior Knowledge Assisted Image Classification using Graph Convolution Networks
• Feature Binding with Category-Dependant MixUp for Semantic Segmentation and Adversarial Robustness
• GuidedMix-Net: Learning to Improve Pseudo Masks Using Labeled Images as Reference
• EventDrop: data augmentation for event-based learning
• SiameseDuo++: Active Learning from Data Streams with Dual Augmented Siamese Networks
• Revisiting data augmentation for subspace clustering
• Data augmentation with Mobius transformations
• Why Mixup Improves the Model Performance
• Direct Differentiable Augmentation Search
• Improved Image Augmentation for Convolutional Neural Networks by Copyout and CopyPairing
• Ferrograph image classification
• LA3: Efficient Label-Aware AutoAugment
• A Visual Domain Transfer Learning Approach for Heartbeat Sound Classification
• Augmented Data as an Auxiliary Plug-in Towards Categorization of Crowdsourced Heritage Data
• SapAugment: Learning A Sample Adaptive Policy for Data Augmentation
• Colonoscopy Polyp Detection: Domain Adaptation From Medical Report Images to Real-time Videos
• A Baseline for Multi-Label Image Classification Using An Ensemble of Deep Convolutional Neural Networks
• Similarity Transfer for Knowledge Distillation
• ScaleNet: An Unsupervised Representation Learning Method for Limited Information
• Understanding and Enhancing the Use of Context for Machine Translation
• On Data-centric Myths
• Unified Multifaceted Feature Learning for Person Re-Identification
• Hierarchical Auxiliary Learning