Cost-Effective Active Learning for Deep Image Classification

Abstract

Recent successes in learning-based image classification, however, heavily rely on the large number of annotated training samples, which may require considerable human efforts. In this paper, we propose a novel active learning framework, which is capable of building a competitive classifier with optimal feature representation via a limited amount of labeled training instances in an incremental learning manner. Our approach advances the existing active learning methods in two aspects. First, we incorporate deep convolutional neural networks into active learning. Through the properly designed framework, the feature representation and the classifier can be simultaneously updated with progressively annotated informative samples. Second, we present a cost-effective sample selection strategy to improve the classification performance with less manual annotations. Unlike traditional methods focusing on only the uncertain samples of low prediction confidence, we especially discover the large amount of high confidence samples from the unlabeled set for feature learning. Specifically, these high confidence samples are automatically selected and iteratively assigned pseudo-labels. We thus call our framework "Cost-Effective Active Learning" (CEAL) standing for the two advantages.Extensive experiments demonstrate that the proposed CEAL framework can achieve promising results on two challenging image classification datasets, i.e., face recognition on CACD database [1] and object categorization on Caltech-256 [2].

Accepted by IEEE Transactions on Circuits and Systems for Video Technology (TCSVT) 2016

References in corpus (2)

• Very Deep Convolutional Networks for Large-Scale Image Recognition
• DeCAF: A Deep Convolutional Activation Feature for Generic Visual Recognition

Cited by in corpus (73)

• Active Learning for Convolutional Neural Networks: A Core-Set Approach
• A Survey on Active Learning and Human-in-the-Loop Deep Learning for Medical Image Analysis
• T-CNN: Tubelets with Convolutional Neural Networks for Object Detection from Videos
• BatchBALD: Efficient and Diverse Batch Acquisition for Deep Bayesian Active Learning
• Self-Training for End-to-End Speech Recognition
• Pervasive AI for IoT applications: A Survey on Resource-efficient Distributed Artificial Intelligence
• Active Self-Paced Learning for Cost-Effective and Progressive Face Identification
• A Benchmark and Comparison of Active Learning for Logistic Regression
• Optical Remote Sensing Image Understanding with Weak Supervision: Concepts, Methods, and Perspectives
• Active Learning with Point Supervision for Cost-Effective Panicle Detection in Cereal Crops
• AutoTune: Automatically Tuning Convolutional Neural Networks for Improved Transfer Learning
• Computer Vision with Deep Learning for Plant Phenotyping in Agriculture: A Survey
• Active Learning with Partial Feedback
• Unsupervised Intra-domain Adaptation for Semantic Segmentation through Self-Supervision
• Trading-off Accuracy and Energy of Deep Inference on Embedded Systems: A Co-Design Approach
• On Statistical Bias In Active Learning: How and When To Fix It
• A Survey on Deep Learning of Small Sample in Biomedical Image Analysis
• Parting with Illusions about Deep Active Learning
• GCT: Graph Co-Training for Semi-Supervised Few-Shot Learning
• Deep Active Learning for Computer Vision: Past and Future
• Reinforced active learning for image segmentation
• Deep Learning Body Region Classification of MRI and CT examinations
• Replay in Deep Learning: Current Approaches and Missing Biological Elements
• Active and Incremental Learning with Weak Supervision
• Deep Active Learning in Remote Sensing for data efficient Change Detection
• CheXpert++: Approximating the CheXpert labeler for Speed,Differentiability, and Probabilistic Output
• Weight-dependent Gates for Network Pruning
• Adversarial Representation Active Learning
• A Self Supervised StyleGAN for Image Annotation and Classification with Extremely Limited Labels
• Low-resource Deep Entity Resolution with Transfer and Active Learning
• Minimax Active Learning
• Deeply Supervised Active Learning for Finger Bones Segmentation
• Deep Reinforcement One-Shot Learning for Artificially Intelligent Classification Systems
• Minority Class Oriented Active Learning for Imbalanced Datasets
• Active Learning to Classify Macromolecular Structures in situ for Less Supervision in Cryo-Electron Tomography
• Active Learning for Entity Alignment
• Semi-supervised Active Learning for Instance Segmentation via Scoring Predictions
• Deep Co-Space: Sample Mining Across Feature Transformation for Semi-Supervised Learning
• Temporal Output Discrepancy for Loss Estimation-based Active Learning
• Active Learning with TensorBoard Projector
• Active learning using weakly supervised signals for quality inspection
• Diffusion-based Deep Active Learning
• Multiple instance active learning for object detection
• Learning to Sample the Most Useful Training Patches from Images
• Human-in-the-loop Extraction of Interpretable Concepts in Deep Learning Models
• DEAL: Difficulty-aware Active Learning for Semantic Segmentation
• One-bit Supervision for Image Classification
• Weakly Supervised Active Learning with Cluster Annotation
• ReDAL: Region-based and Diversity-aware Active Learning for Point Cloud Semantic Segmentation
• Exquisitor: Interactive Learning at Large
• Improved Synthetic Training for Reading Comprehension
• One-bit Supervision for Image Classification: Problem, Solution, and Beyond
• On the Robustness of Active Learning
• Confident Coreset for Active Learning in Medical Image Analysis
• Autonomous Curiosity for Real-Time Training Onboard Robotic Agents
• On-the-Fly Point Annotation for Fast Medical Video Labeling
• Deep Multi-Scale Features Learning for Distorted Image Quality Assessment
• Integrating Crowdsourcing and Active Learning for Classification of Work-Life Events from Tweets
• Just Label What You Need: Fine-Grained Active Selection for Perception and Prediction through Partially Labeled Scenes
• Pseudo-Labeling for Small Lesion Detection on Diabetic Retinopathy Images
• Augmented Memory Networks for Streaming-Based Active One-Shot Learning
• Minimizing Supervision in Multi-label Categorization
• Active Crowd Counting with Limited Supervision
• h-calibration: Rethinking Classifier Recalibration with Probabilistic Error-Bounded Objective
• Active Learning for UAV-based Semantic Mapping
• Distance in Latent Space as Novelty Measure
• PipeSim: Trace-driven Simulation of Large-Scale AI Operations Platforms
• Informative sample generation using class aware generative adversarial networks for classification of chest Xrays
• Loss Prediction: End-to-End Active Learning Approach For Speech Recognition
• Active learning for medical code assignment
• Person Re-Identification via Active Hard Sample Mining
• MDFM: Multi-Decision Fusing Model for Few-Shot Learning
• Dataset Culling: Towards Efficient Training Of Distillation-Based Domain Specific Models