Data Augmentation using Random Image Cropping and Patching for Deep CNNs

Abstract

Deep convolutional neural networks (CNNs) have achieved remarkable results in image processing tasks. However, their high expression ability risks overfitting. Consequently, data augmentation techniques have been proposed to prevent overfitting while enriching datasets. Recent CNN architectures with more parameters are rendering traditional data augmentation techniques insufficient. In this study, we propose a new data augmentation technique called random image cropping and patching (RICAP) which randomly crops four images and patches them to create a new training image. Moreover, RICAP mixes the class labels of the four images, resulting in an advantage similar to label smoothing. We evaluated RICAP with current state-of-the-art CNNs (e.g., the shake-shake regularization model) by comparison with competitive data augmentation techniques such as cutout and mixup. RICAP achieves a new state-of-the-art test error of $2.19\%$ on CIFAR-10. We also confirmed that deep CNNs with RICAP achieve better results on classification tasks using CIFAR-100 and ImageNet and an image-caption retrieval task using Microsoft COCO.

accepted version, 16 pages

References in corpus (16)

• Very Deep Convolutional Networks for Large-Scale Image Recognition
• Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift
• Distilling the Knowledge in a Neural Network
• Explaining and Harnessing Adversarial Examples
• Improving neural networks by preventing co-adaptation of feature detectors
• Neural Architecture Search with Reinforcement Learning
• Improved Regularization of Convolutional Neural Networks with Cutout
• Striving for Simplicity: The All Convolutional Net
• FitNets: Hints for Thin Deep Nets
• Unifying Visual-Semantic Embeddings with Multimodal Neural Language Models
• Person Re-identification: Past, Present and Future
• Random Erasing Data Augmentation
• AutoAugment: Learning Augmentation Policies from Data
• Visualizing Deep Neural Network Decisions: Prediction Difference Analysis
• Shake-Shake regularization
• Revisiting Unreasonable Effectiveness of Data in Deep Learning Era

Cited by in corpus (34)

• AugMix: A Simple Data Processing Method to Improve Robustness and Uncertainty
• Single Model Deep Learning on Imbalanced Small Datasets for Skin Lesion Classification
• Automatic classification between COVID-19 pneumonia, non-COVID-19 pneumonia, and the healthy on chest X-ray image: combination of data augmentation methods
• BDANet: Multiscale Convolutional Neural Network with Cross-directional Attention for Building Damage Assessment from Satellite Images
• FMix: Enhancing Mixed Sample Data Augmentation
• A Survey of Deep Learning for Scientific Discovery
• Data Augmentation for Object Detection via Progressive and Selective Instance-Switching
• Physical Knowledge Enhanced Deep Neural Network for Sea Surface Temperature Prediction
• Automated Defect Recognition of Castings defects using Neural Networks
• Increasing the Robustness of Semantic Segmentation Models with Painting-by-Numbers
• Artificial and beneficial -- Exploiting artificial images for aerial vehicle detection
• Age Range Estimation using MTCNN and VGG-Face Model
• SnapMix: Semantically Proportional Mixing for Augmenting Fine-grained Data
• Measuring Visual Generalization in Continuous Control from Pixels
• MSE Loss with Outlying Label for Imbalanced Classification
• Traffic Cameras to detect inland waterway barge traffic: An Application of machine learning
• KinePose: A temporally optimized inverse kinematics technique for 6DOF human pose estimation with biomechanical constraints
• PK-GCN: Prior Knowledge Assisted Image Classification using Graph Convolution Networks
• Synthetic Data Generation for Augmenting Small Samples
• Student Network Learning via Evolutionary Knowledge Distillation
• Random Shadows and Highlights: A new data augmentation method for extreme lighting conditions
• StackMix: A complementary Mix algorithm
• Data augmentation with Mobius transformations
• Split Unlearning
• Self-paced Data Augmentation for Training Neural Networks
• Untapped Potential of Data Augmentation: A Domain Generalization Viewpoint
• Discriminative Cross-Modal Data Augmentation for Medical Imaging Applications
• Autonomous Driving Implementation in an Experimental Environment
• Isotonic Data Augmentation for Knowledge Distillation
• Training Lightweight CNNs for Human-Nanodrone Proximity Interaction from Small Datasets using Background Randomization
• Using Fictitious Class Representations to Boost Discriminative Zero-Shot Learners
• Survey on Deep Learning-based Kuzushiji Recognition
• Domain-Agnostic Clustering with Self-Distillation
• Data augmentation by morphological mixup for solving Raven's Progressive Matrices