Places: An Image Database for Deep Scene Understanding

Abstract

The rise of multi-million-item dataset initiatives has enabled data-hungry machine learning algorithms to reach near-human semantic classification at tasks such as object and scene recognition. Here we describe the Places Database, a repository of 10 million scene photographs, labeled with scene semantic categories and attributes, comprising a quasi-exhaustive list of the types of environments encountered in the world. Using state of the art Convolutional Neural Networks, we provide impressive baseline performances at scene classification. With its high-coverage and high-diversity of exemplars, the Places Database offers an ecosystem to guide future progress on currently intractable visual recognition problems.

References in corpus (5)

• Very Deep Convolutional Networks for Large-Scale Image Recognition
• DeCAF: A Deep Convolutional Activation Feature for Generic Visual Recognition
• The Cityscapes Dataset for Semantic Urban Scene Understanding
• Synthesizing the preferred inputs for neurons in neural networks via deep generator networks
• Semantic Understanding of Scenes through the ADE20K Dataset

Cited by in corpus (52)

• Building Instance Classification Using Street View Images
• Wider or Deeper: Revisiting the ResNet Model for Visual Recognition
• Context Based Emotion Recognition using EMOTIC Dataset
• Self-supervised Visual Feature Learning with Deep Neural Networks: A Survey
• Knowledge Guided Disambiguation for Large-Scale Scene Classification with Multi-Resolution CNNs
• Accelerating 3D Deep Learning with PyTorch3D
• An Ensemble of Convolutional Neural Networks for Audio Classification
• Detecting Visual Relationships with Deep Relational Networks
• Analysis and Optimization of Convolutional Neural Network Architectures
• Lightweight Modules for Efficient Deep Learning based Image Restoration
• Learning Deep NBNN Representations for Robust Place Categorization
• Scene Image Representation by Foreground, Background and Hybrid Features
• Sharing Residual Units Through Collective Tensor Factorization in Deep Neural Networks
• Content and Context Features for Scene Image Representation
• Image Inpainting using Block-wise Procedural Training with Annealed Adversarial Counterpart
• Situation Recognition with Graph Neural Networks
• Global Model Interpretation via Recursive Partitioning
• HDF: Hybrid Deep Features for Scene Image Representation
• Multi-modal panoramic 3D outdoor datasets for place categorization
• Leveraging Semantic Scene Characteristics and Multi-Stream Convolutional Architectures in a Contextual Approach for Video-Based Visual Emotion Recognition in the Wild
• Why my photos look sideways or upside down? Detecting Canonical Orientation of Images using Convolutional Neural Networks
• WebVision Challenge: Visual Learning and Understanding With Web Data
• Active Convolution: Learning the Shape of Convolution for Image Classification
• Deep Temporal Linear Encoding Networks
• To Balance or Not to Balance: A Simple-yet-Effective Approach for Learning with Long-Tailed Distributions
• Relating Input Concepts to Convolutional Neural Network Decisions
• Tag-based Semantic Features for Scene Image Classification
• Affordance Learning In Direct Perception for Autonomous Driving
• Neural Architecture Search for Deep Image Prior
• An Out-of-the-box Full-network Embedding for Convolutional Neural Networks
• Residual Squeeze VGG16
• Self-Supervised Visual Place Recognition Learning in Mobile Robots
• Monocular Depth Estimation with Augmented Ordinal Depth Relationships
• Pose-GNN : Camera Pose Estimation System Using Graph Neural Networks
• Cross-Domain Self-supervised Multi-task Feature Learning using Synthetic Imagery
• Learning from Noisy Web Data with Category-level Supervision
• Multi-modal Geolocation Estimation Using Deep Neural Networks
• On the Behavior of Convolutional Nets for Feature Extraction
• SLIDE: Single Image 3D Photography with Soft Layering and Depth-aware Inpainting
• Unsupervised Multi-label Dataset Generation from Web Data
• AV1 Video Coding Using Texture Analysis With Convolutional Neural Networks
• Hierarchical approach to classify food scenes in egocentric photo-streams
• Urban land-use analysis using proximate sensing imagery: a survey
• Non-iterative recomputation of dense layers for performance improvement of DCNN
• Fine-grained Classification using Heterogeneous Web Data and Auxiliary Categories
• Fast Fourier Color Constancy and Grayness Index for ISPA Illumination Estimation Challenge
• The Compressed Model of Residual CNDS
• Residual-CNDS for Grand Challenge Scene Dataset
• Generating Video Descriptions with Topic Guidance
• Salvage Reusable Samples from Noisy Data for Robust Learning
• Bayesian deep learning of affordances from RGB images
• Convolutional Neural Networks Based Texture Modeling For AV1