Deep Reinforcement Learning in Computer Vision: A Comprehensive Survey

Abstract

Deep reinforcement learning augments the reinforcement learning framework and utilizes the powerful representation of deep neural networks. Recent works have demonstrated the remarkable successes of deep reinforcement learning in various domains including finance, medicine, healthcare, video games, robotics, and computer vision. In this work, we provide a detailed review of recent and state-of-the-art research advances of deep reinforcement learning in computer vision. We start with comprehending the theories of deep learning, reinforcement learning, and deep reinforcement learning. We then propose a categorization of deep reinforcement learning methodologies and discuss their advantages and limitations. In particular, we divide deep reinforcement learning into seven main categories according to their applications in computer vision, i.e. (i)landmark localization (ii) object detection; (iii) object tracking; (iv) registration on both 2D image and 3D image volumetric data (v) image segmentation; (vi) videos analysis; and (vii) other applications. Each of these categories is further analyzed with reinforcement learning techniques, network design, and performance. Moreover, we provide a comprehensive analysis of the existing publicly available datasets and examine source code availability. Finally, we present some open issues and discuss future research directions on deep reinforcement learning in computer vision

References in corpus (58)

• Very Deep Convolutional Networks for Large-Scale Image Recognition
• Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks
• Continuous control with deep reinforcement learning
• YOLOv3: An Incremental Improvement
• High-Speed Tracking with Kernelized Correlation Filters
• DeepPose: Human Pose Estimation via Deep Neural Networks
• Trust Region Policy Optimization
• Efficient Inference in Fully Connected CRFs with Gaussian Edge Potentials
• A Contextual-Bandit Approach to Personalized News Article Recommendation
• GOT-10k: A Large High-Diversity Benchmark for Generic Object Tracking in the Wild
• Attention-Based Models for Speech Recognition
• MOT16: A Benchmark for Multi-Object Tracking
• REFUGE Challenge: A Unified Framework for Evaluating Automated Methods for Glaucoma Assessment from Fundus Photographs
• Deep Bilateral Learning for Real-Time Image Enhancement
• MOTChallenge 2015: Towards a Benchmark for Multi-Target Tracking
• The Goldilocks Principle: Reading Children's Books with Explicit Memory Representations
• DeepIGeoS: A Deep Interactive Geodesic Framework for Medical Image Segmentation
• Learning to reinforcement learn
• Safe, Multi-Agent, Reinforcement Learning for Autonomous Driving
• SfM-Net: Learning of Structure and Motion from Video
• Meta-World: A Benchmark and Evaluation for Multi-Task and Meta Reinforcement Learning
• CARLA: An Open Urban Driving Simulator
• Multi-agent Reinforcement Learning in Sequential Social Dilemmas
• Visual Saliency Based on Multiscale Deep Features
• Benchmarking Model-Based Reinforcement Learning
• Efficient Off-Policy Meta-Reinforcement Learning via Probabilistic Context Variables
• A Hierarchical Neural Autoencoder for Paragraphs and Documents
• A Survey of Learning in Multiagent Environments: Dealing with Non-Stationarity
• FusionSeg: Learning to combine motion and appearance for fully automatic segmention of generic objects in videos
• Synthesizing Programs for Images using Reinforced Adversarial Learning
• Reinforcement Learning in Healthcare: A Survey
• Model-Based Reinforcement Learning via Meta-Policy Optimization
• Meta Reinforcement Learning with Latent Variable Gaussian Processes
• Deep Reinforcement Learning for Visual Object Tracking in Videos
• Sim-to-Real Reinforcement Learning for Deformable Object Manipulation
• Hierarchical Object Detection with Deep Reinforcement Learning
• Tree-Structured Reinforcement Learning for Sequential Object Localization
• Deep Reinforcement Learning for Unsupervised Video Summarization with Diversity-Representativeness Reward
• Fast YOLO: A Fast You Only Look Once System for Real-time Embedded Object Detection in Video
• On the Effects of Batch and Weight Normalization in Generative Adversarial Networks
• Fast Multiple Landmark Localisation Using a Patch-based Iterative Network
• SRM : A Style-based Recalibration Module for Convolutional Neural Networks
• Learning to Paint With Model-based Deep Reinforcement Learning
• Crowd Counting using Deep Recurrent Spatial-Aware Network
• Robust Multimodal Image Registration Using Deep Recurrent Reinforcement Learning
• Experience Replay Optimization
• Multi-step Reinforcement Learning: A Unifying Algorithm
• Multi-modal Visual Tracking: Review and Experimental Comparison
• Deep Reinforcement Learning with Model Learning and Monte Carlo Tree Search in Minecraft
• Generic Object Detection With Dense Neural Patterns and Regionlets
• Reinforcement Learning for Semantic Segmentation in Indoor Scenes
• A Real-Time Model-Based Reinforcement Learning Architecture for Robot Control
• Deep Model-Based Reinforcement Learning for High-Dimensional Problems, a Survey
• Invertible Residual Network with Regularization for Effective Medical Image Segmentation
• Reinforcement Learning-based Automatic Diagnosis of Acute Appendicitis in Abdominal CT
• Ultrasound-Guided Robotic Navigation with Deep Reinforcement Learning
• Ultrasound Video Summarization using Deep Reinforcement Learning
• Offset Curves Loss for Imbalanced Problem in Medical Segmentation