A Simple Semi-Supervised Learning Framework for Object Detection

Abstract

Semi-supervised learning (SSL) has a potential to improve the predictive performance of machine learning models using unlabeled data. Although there has been remarkable recent progress, the scope of demonstration in SSL has mainly been on image classification tasks. In this paper, we propose STAC, a simple yet effective SSL framework for visual object detection along with a data augmentation strategy. STAC deploys highly confident pseudo labels of localized objects from an unlabeled image and updates the model by enforcing consistency via strong augmentations. We propose experimental protocols to evaluate the performance of semi-supervised object detection using MS-COCO and show the efficacy of STAC on both MS-COCO and VOC07. On VOC07, STAC improves the AP${}^{0.5}$ from $76.30$ to $79.08$; on MS-COCO, STAC demonstrates $2\times$ higher data efficiency by achieving 24.38 mAP using only 5\% labeled data than supervised baseline that marks 23.86\% using 10\% labeled data. The code is available at https://github.com/google-research/ssl_detection/.

References in corpus (17)

• Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks
• mixup: Beyond Empirical Risk Minimization
• R-FCN: Object Detection via Region-based Fully Convolutional Networks
• Improved Regularization of Convolutional Neural Networks with Cutout
• FixMatch: Simplifying Semi-Supervised Learning with Consistency and Confidence
• Unsupervised Data Augmentation for Consistency Training
• Mean teachers are better role models: Weight-averaged consistency targets improve semi-supervised deep learning results
• Temporal Ensembling for Semi-Supervised Learning
• Random Erasing Data Augmentation
• Semi-Supervised Learning with Ladder Networks
• AugMix: A Simple Data Processing Method to Improve Robustness and Uncertainty
• EfficientDet: Scalable and Efficient Object Detection
• Learning with Pseudo-Ensembles
• Self-training with Noisy Student improves ImageNet classification
• Bag of Freebies for Training Object Detection Neural Networks
• SpineNet: Learning Scale-Permuted Backbone for Recognition and Localization
• Proposal Learning for Semi-Supervised Object Detection

Cited by in corpus (47)

• Rethinking Pre-training and Self-training
• PseudoSeg: Designing Pseudo Labels for Semantic Segmentation
• MutexMatch: Semi-Supervised Learning with Mutex-Based Consistency Regularization
• Semi-Supervised Training to Improve Player and Ball Detection in Soccer
• SemiFL: Semi-Supervised Federated Learning for Unlabeled Clients with Alternate Training
• SODA10M: A Large-Scale 2D Self/Semi-Supervised Object Detection Dataset for Autonomous Driving
• Unbiased Teacher for Semi-Supervised Object Detection
• Instant-Teaching: An End-to-End Semi-Supervised Object Detection Framework
• Propagate Yourself: Exploring Pixel-Level Consistency for Unsupervised Visual Representation Learning
• Poisoning the Unlabeled Dataset of Semi-Supervised Learning
• ST++: Make Self-training Work Better for Semi-supervised Semantic Segmentation
• Semi-supervised 3D Object Detection via Adaptive Pseudo-Labeling
• PseudoAugment: Learning to Use Unlabeled Data for Data Augmentation in Point Clouds
• Artificial and beneficial -- Exploiting artificial images for aerial vehicle detection
• A Comparative Review of Recent Few-Shot Object Detection Algorithms
• Adaptive Self-Training for Object Detection
• Consistency-based Semi-supervised Active Learning: Towards Minimizing Labeling Cost
• Narrowing the Gap: Improved Detector Training with Noisy Location Annotations
• 3DIoUMatch: Leveraging IoU Prediction for Semi-Supervised 3D Object Detection
• Multimodal Semi-Supervised Learning for 3D Objects
• Predicting Class Distribution Shift for Reliable Domain Adaptive Object Detection
• Augmentation Strategies for Learning with Noisy Labels
• Combating Noise: Semi-supervised Learning by Region Uncertainty Quantification
• Learning Fast Sample Re-weighting Without Reward Data
• SST: Self-training with Self-adaptive Thresholding for Semi-supervised Learning
• Cycle Self-Training for Semi-Supervised Object Detection with Distribution Consistency Reweighting
• Reliable Student: Addressing Noise in Semi-Supervised 3D Object Detection
• Revisiting Knowledge Distillation for Object Detection
• Points as Queries: Weakly Semi-supervised Object Detection by Points
• WSSOD: A New Pipeline for Weakly- and Semi-Supervised Object Detection
• Towards Few-Annotation Learning for Object Detection: Are Transformer-based Models More Efficient ?
• Pseudo-labeling for Scalable 3D Object Detection
• EBJR: Energy-Based Joint Reasoning for Adaptive Inference
• Lesion Localization in OCT by Semi-Supervised Object Detection
• Humble Teachers Teach Better Students for Semi-Supervised Object Detection
• SSMD: Semi-Supervised Medical Image Detection with Adaptive Consistency and Heterogeneous Perturbation
• MosaicOS: A Simple and Effective Use of Object-Centric Images for Long-Tailed Object Detection
• Pseudo-Labeling Enhanced by Privileged Information and Its Application to In Situ Sequencing Images
• Semi-supervised Cell Detection in Time-lapse Images Using Temporal Consistency
• Improving Localization for Semi-Supervised Object Detection
• Non-iterative optimization of pseudo-labeling thresholds for training object detection models from multiple datasets
• Hetero-Modal Learning and Expansive Consistency Constraints for Semi-Supervised Detection from Multi-Sequence Data
• VeniBot: Towards Autonomous Venipuncture with Semi-supervised Vein Segmentation from Ultrasound Images
• Semi-supervised Learning for Dense Object Detection in Retail Scenes
• Contrast R-CNN for Continual Learning in Object Detection
• Scalable Semi-supervised Landmark Localization for X-ray Images using Few-shot Deep Adaptive Graph
• Bootstrap Your Object Detector via Mixed Training