DeCAF: A Deep Convolutional Activation Feature for Generic Visual Recognition

Abstract

We evaluate whether features extracted from the activation of a deep convolutional network trained in a fully supervised fashion on a large, fixed set of object recognition tasks can be re-purposed to novel generic tasks. Our generic tasks may differ significantly from the originally trained tasks and there may be insufficient labeled or unlabeled data to conventionally train or adapt a deep architecture to the new tasks. We investigate and visualize the semantic clustering of deep convolutional features with respect to a variety of such tasks, including scene recognition, domain adaptation, and fine-grained recognition challenges. We compare the efficacy of relying on various network levels to define a fixed feature, and report novel results that significantly outperform the state-of-the-art on several important vision challenges. We are releasing DeCAF, an open-source implementation of these deep convolutional activation features, along with all associated network parameters to enable vision researchers to be able to conduct experimentation with deep representations across a range of visual concept learning paradigms.

References in corpus (3)

• Improving neural networks by preventing co-adaptation of feature detectors
• Frustratingly Easy Domain Adaptation
• ImageNet Large Scale Visual Recognition Challenge

Cited by in corpus (692)

• Deep Learning in Neural Networks: An Overview
• A Simple Framework for Contrastive Learning of Visual Representations
• Model-Agnostic Meta-Learning for Fast Adaptation of Deep Networks
• Caffe: Convolutional Architecture for Fast Feature Embedding
• How transferable are features in deep neural networks?
• Bootstrap your own latent: A new approach to self-supervised Learning
• Brain Tumor Segmentation with Deep Neural Networks
• Spatial Pyramid Pooling in Deep Convolutional Networks for Visual Recognition
• Learning Transferable Features with Deep Adaptation Networks
• Unsupervised Domain Adaptation by Backpropagation
• Accurate, Large Minibatch SGD: Training ImageNet in 1 Hour
• Deep Domain Confusion: Maximizing for Domain Invariance
• Deep Visual Domain Adaptation: A Survey
• DeepDTA: Deep Drug-Target Binding Affinity Prediction
• SGDR: Stochastic Gradient Descent with Warm Restarts
• Understanding Black-box Predictions via Influence Functions
• Deep Subdomain Adaptation Network for Image Classification
• Unsupervised Domain Adaptation with Residual Transfer Networks
• Matching Networks for One Shot Learning
• Show and Tell: Lessons learned from the 2015 MSCOCO Image Captioning Challenge
• Fully Convolutional Networks for Semantic Segmentation
• CNN: Single-label to Multi-label
• Generalizing from a Few Examples: A Survey on Few-Shot Learning
• VL-BERT: Pre-training of Generic Visual-Linguistic Representations
• Adversarial Feature Learning
• Learning Representations by Maximizing Mutual Information Across Views
• Cost-Effective Active Learning for Deep Image Classification
• CNN Features off-the-shelf: an Astounding Baseline for Recognition
• DenseNet: Implementing Efficient ConvNet Descriptor Pyramids
• DeepSaliency: Multi-Task Deep Neural Network Model for Salient Object Detection
• Compressing Neural Networks with the Hashing Trick
• Transductive Multi-view Zero-Shot Learning
• Online Tracking by Learning Discriminative Saliency Map with Convolutional Neural Network
• Beyond Sharing Weights for Deep Domain Adaptation
• Land Use Classification in Remote Sensing Images by Convolutional Neural Networks
• Advanced Steel Microstructural Classification by Deep Learning Methods
• Convolutional Feature Masking for Joint Object and Stuff Segmentation
• Adversarial Discriminative Domain Adaptation
• Generalisation in humans and deep neural networks
• Visual Saliency Detection Based on Multiscale Deep CNN Features
• Bird Species Categorization Using Pose Normalized Deep Convolutional Nets
• Aligning Domain-specific Distribution and Classifier for Cross-domain Classification from Multiple Sources
• Rethinking Pre-training and Self-training
• Aggregated Residual Transformations for Deep Neural Networks
• Revisiting Batch Normalization For Practical Domain Adaptation
• Explain Images with Multimodal Recurrent Neural Networks
• Deep CORAL: Correlation Alignment for Deep Domain Adaptation
• Spectral Signatures in Backdoor Attacks
• Person Re-Identification by Camera Correlation Aware Feature Augmentation
• The Extreme Value Machine
• Supervised Learning of Semantics-Preserving Hash via Deep Convolutional Neural Networks
• Recent Advances in Convolutional Neural Networks
• What makes ImageNet good for transfer learning?
• Fairness in Machine Learning
• Domain Adaptation for Visual Applications: A Comprehensive Survey
• Multi-Representation Adaptation Network for Cross-domain Image Classification
• Tent: Fully Test-time Adaptation by Entropy Minimization
• Deep Convolutional Ranking for Multilabel Image Annotation
• Visual Saliency Based on Multiscale Deep Features
• A Primer on Motion Capture with Deep Learning: Principles, Pitfalls and Perspectives
• Descriptor Matching with Convolutional Neural Networks: a Comparison to SIFT
• Deep Learning for Generic Object Detection: A Survey
• Deep Gaze I: Boosting Saliency Prediction with Feature Maps Trained on ImageNet
• A Survey on Content-Aware Video Analysis for Sports
• Contextual Encoder-Decoder Network for Visual Saliency Prediction
• Improving Object Detection with Deep Convolutional Networks via Bayesian Optimization and Structured Prediction
• One-Shot Imitation Learning
• Do Convolutional Neural Networks Learn Class Hierarchy?
• Looking Beyond Appearances: Synthetic Training Data for Deep CNNs in Re-identification
• Learning Deep Features for Discriminative Localization
• Realistic Evaluation of Deep Semi-Supervised Learning Algorithms
• Convolutional Neural Network-based Place Recognition
• Weakly Supervised Fine-Grained Image Categorization
• Deep Aesthetic Quality Assessment with Semantic Information
• Simultaneous Detection and Segmentation
• LSDA: Large Scale Detection Through Adaptation
• Effective Multi-Query Expansions: Collaborative Deep Networks for Robust Landmark Retrieval
• Deep Learning Techniques for Future Intelligent Cross-Media Retrieval
• Momentum Contrastive Learning for Few-Shot COVID-19 Diagnosis from Chest CT Images
• Anatomy-specific classification of medical images using deep convolutional nets
• A Richly Annotated Dataset for Pedestrian Attribute Recognition
• Towards Robust Evaluations of Continual Learning
• Hybrid CNN and Dictionary-Based Models for Scene Recognition and Domain Adaptation
• Places: An Image Database for Deep Scene Understanding
• Convolutional Kernel Networks
• Zero-Shot Learning -- A Comprehensive Evaluation of the Good, the Bad and the Ugly
• Online Batch Selection for Faster Training of Neural Networks
• Deep Residual Correction Network for Partial Domain Adaptation
• Simultaneous Traffic Sign Detection and Boundary Estimation using Convolutional Neural Network
• Do Convnets Learn Correspondence?
• Automated Machine Learning: State-of-The-Art and Open Challenges
• Third-Person Imitation Learning
• Learning Deep Representations of Fine-grained Visual Descriptions
• The Devil is in the Tails: Fine-grained Classification in the Wild
• On the Performance of ConvNet Features for Place Recognition
• Predicting Deep Zero-Shot Convolutional Neural Networks using Textual Descriptions
• Semantic-Aware Scene Recognition
• R-CNNs for Pose Estimation and Action Detection
• Learning Contextual Dependencies with Convolutional Hierarchical Recurrent Neural Networks
• Learning to cluster in order to transfer across domains and tasks
• Visual Sentiment Prediction with Deep Convolutional Neural Networks
• Deep Transfer Learning with Joint Adaptation Networks
• SuperAnimal pretrained pose estimation models for behavioral analysis
• Transformer with Peak Suppression and Knowledge Guidance for Fine-grained Image Recognition
• Do Adversarially Robust ImageNet Models Transfer Better?
• Return of Frustratingly Easy Domain Adaptation
• DeXpression: Deep Convolutional Neural Network for Expression Recognition
• Unsupervised Domain Adaptation with Adversarial Residual Transform Networks
• Training Convolutional Neural Networks with Limited Training Data for Ear Recognition in the Wild
• Weakly-supervised Discovery of Visual Pattern Configurations
• Decoding and mapping task states of the human brain via deep learning
• FoodAI: Food Image Recognition via Deep Learning for Smart Food Logging
• Deep Learning Face Attributes in the Wild
• Approaching the Computational Color Constancy as a Classification Problem through Deep Learning
• Rethinking ImageNet Pre-training
• Machine Learning for Scent: Learning Generalizable Perceptual Representations of Small Molecules
• Generative Image Modeling Using Spatial LSTMs
• Locally-Supervised Deep Hybrid Model for Scene Recognition
• DELTA: DEep Learning Transfer using Feature Map with Attention for Convolutional Networks
• Supervised Hashing Using Graph Cuts and Boosted Decision Trees
• A Unified Perspective on Multi-Domain and Multi-Task Learning
• Adversarial Examples for Semantic Segmentation and Object Detection
• Understanding Image Virality
• Wasserstein Discriminant Analysis
• Kernel Manifold Alignment
• From Captions to Visual Concepts and Back
• Multi-Adversarial Domain Adaptation
• Forecasting with time series imaging
• Flowing ConvNets for Human Pose Estimation in Videos
• How Transferable are Neural Networks in NLP Applications?
• Bridging Saliency Detection to Weakly Supervised Object Detection Based on Self-paced Curriculum Learning
• Would Mega-scale Datasets Further Enhance Spatiotemporal 3D CNNs?
• Neural Activation Constellations: Unsupervised Part Model Discovery with Convolutional Networks
• Video Summarization with Long Short-term Memory
• Learning Patterns of Tourist Movement and Photography from Geotagged Photos at Archaeological Heritage Sites in Cuzco, Peru
• DeepID-Net: Deformable Deep Convolutional Neural Networks for Object Detection
• Evaluating Protein Transfer Learning with TAPE
• Multi-label Learning Based Deep Transfer Neural Network for Facial Attribute Classification
• Fully Convolutional Neural Networks for Page Segmentation of Historical Document Images
• Mind the Class Weight Bias: Weighted Maximum Mean Discrepancy for Unsupervised Domain Adaptation
• SpeechStew: Simply Mix All Available Speech Recognition Data to Train One Large Neural Network
• Generic 3D Representation via Pose Estimation and Matching
• Feature-Critic Networks for Heterogeneous Domain Generalization
• LEEP: A New Measure to Evaluate Transferability of Learned Representations
• Do Convolutional Networks need to be Deep for Text Classification ?
• GPT-GNN: Generative Pre-Training of Graph Neural Networks
• Analyzing the Performance of Multilayer Neural Networks for Object Recognition
• A Survey of Deep Network Solutions for Learning Control in Robotics: From Reinforcement to Imitation
• GPU Asynchronous Stochastic Gradient Descent to Speed Up Neural Network Training
• VBPR: Visual Bayesian Personalized Ranking from Implicit Feedback
• HashNet: Deep Learning to Hash by Continuation
• End-to-End Deep Learning for Steering Autonomous Vehicles Considering Temporal Dependencies
• Fine-grained Visual-textual Representation Learning
• Generalized Domain Conditioned Adaptation Network
• Joint Distribution Optimal Transportation for Domain Adaptation
• Encoding High Dimensional Local Features by Sparse Coding Based Fisher Vectors
• One-Shot Adaptation of Supervised Deep Convolutional Models
• Adversarial Robustness as a Prior for Learned Representations
• Significance of Softmax-based Features in Comparison to Distance Metric Learning-based Features
• Knowledge Distillation in Generations: More Tolerant Teachers Educate Better Students
• On the Theory of Transfer Learning: The Importance of Task Diversity
• Sample-to-Sample Correspondence for Unsupervised Domain Adaptation
• Partial Transfer Learning with Selective Adversarial Networks
• Rethinking the Hyperparameters for Fine-tuning
• Learning like a Child: Fast Novel Visual Concept Learning from Sentence Descriptions of Images
• Deep Manifold Traversal: Changing Labels with Convolutional Features
• A Probabilistic Quality Representation Approach to Deep Blind Image Quality Prediction
• Learning Transferable Parameters for Unsupervised Domain Adaptation
• Toward quantitative fractography using convolutional neural networks
• DeepLogo: Hitting Logo Recognition with the Deep Neural Network Hammer
• Deep Reconstruction-Classification Networks for Unsupervised Domain Adaptation
• Weakly Supervised Object Detection in Artworks
• Label Efficient Learning of Transferable Representations across Domains and Tasks
• Overcoming Small Minirhizotron Datasets Using Transfer Learning
• On Leveraging Pretrained GANs for Generation with Limited Data
• A comparison of dense region detectors for image search and fine-grained classification
• Fashion Apparel Detection: The Role of Deep Convolutional Neural Network and Pose-dependent Priors
• Visual Understanding via Multi-Feature Shared Learning with Global Consistency
• Deep Learning the City : Quantifying Urban Perception At A Global Scale
• Provable Meta-Learning of Linear Representations
• Fine-grained pose prediction, normalization, and recognition
• Domain Generalization for Object Recognition with Multi-task Autoencoders
• From Images to Sentences through Scene Description Graphs using Commonsense Reasoning and Knowledge
• Domain Generalization via Multidomain Discriminant Analysis
• An exploration of parameter redundancy in deep networks with circulant projections
• On the effectiveness of task granularity for transfer learning
• Synergic Adversarial Label Learning for Grading Retinal Diseases via Knowledge Distillation and Multi-task Learning
• Brain-Inspired Deep Networks for Image Aesthetics Assessment
• Dense and Diverse Capsule Networks: Making the Capsules Learn Better
• K for the Price of 1: Parameter-efficient Multi-task and Transfer Learning
• Are we done with object recognition? The iCub robot's perspective
• SpotTune: Transfer Learning through Adaptive Fine-tuning
• FewRel: A Large-Scale Supervised Few-Shot Relation Classification Dataset with State-of-the-Art Evaluation
• Ridge Regression, Hubness, and Zero-Shot Learning
• Recent Advances in Zero-shot Recognition
• LogME: Practical Assessment of Pre-trained Models for Transfer Learning
• Image Question Answering using Convolutional Neural Network with Dynamic Parameter Prediction
• A Unified Semantic Embedding: Relating Taxonomies and Attributes
• Never Stop Learning: The Effectiveness of Fine-Tuning in Robotic Reinforcement Learning
• Self-supervised Auxiliary Learning with Meta-paths for Heterogeneous Graphs
• Sketch-a-Net that Beats Humans
• Real-time emotion recognition for gaming using deep convolutional network features
• Transfer Learning for Future Wireless Networks: A Comprehensive Survey
• Inferring Latent Domains for Unsupervised Deep Domain Adaptation
• Learning to Transfer Examples for Partial Domain Adaptation
• Part Detector Discovery in Deep Convolutional Neural Networks
• Boosting offline handwritten text recognition in historical documents with few labeled lines
• Multi-Task Incremental Learning for Object Detection
• Deep Layer Aggregation
• Negative Margin Matters: Understanding Margin in Few-shot Classification
• TinyTL: Reduce Activations, Not Trainable Parameters for Efficient On-Device Learning
• Discriminative Unsupervised Feature Learning with Exemplar Convolutional Neural Networks
• Racial Faces in-the-Wild: Reducing Racial Bias by Information Maximization Adaptation Network
• The Unbalanced Gromov Wasserstein Distance: Conic Formulation and Relaxation
• The Lottery Tickets Hypothesis for Supervised and Self-supervised Pre-training in Computer Vision Models
• Transfer Learning with Dynamic Adversarial Adaptation Network
• Towards Effective Codebookless Model for Image Classification
• ViTOR: Learning to Rank Webpages Based on Visual Features
• ElasticTrainer: Speeding Up On-Device Training with Runtime Elastic Tensor Selection
• Rethinking of Pedestrian Attribute Recognition: Realistic Datasets with Efficient Method
• On learning to localize objects with minimal supervision
• Deep Learning for Tactile Understanding From Visual and Haptic Data
• Is it enough to optimize CNN architectures on ImageNet?
• Sparse Deep Stacking Network for Image Classification
• Score Function Features for Discriminative Learning: Matrix and Tensor Framework
• Deep Pyramidal Residual Networks
• Multiple Source Domain Adaptation with Adversarial Training of Neural Networks
• Partial Adversarial Domain Adaptation
• Video Summarization using Deep Semantic Features
• Viewpoints and Keypoints
• How Much Off-The-Shelf Knowledge Is Transferable From Natural Images To Pathology Images?
• Joint Domain Alignment and Discriminative Feature Learning for Unsupervised Deep Domain Adaptation
• Episodic Training for Domain Generalization
• Learning Transferable Policies for Monocular Reactive MAV Control
• Scalable Transfer Learning with Expert Models
• Scaling Video Analytics on Constrained Edge Nodes
• Pixie: A System for Recommending 3+ Billion Items to 200+ Million Users in Real-Time
• Minimum Class Confusion for Versatile Domain Adaptation
• Learning Deep Convolutional Features for MRI Based Alzheimer's Disease Classification
• Unsupervised Extraction of Video Highlights Via Robust Recurrent Auto-encoders
• Transfer Learning for Video Recognition with Scarce Training Data for Deep Convolutional Neural Network
• Evaluating Contrastive Models for Instance-based Image Retrieval
• Pre-Training Graph Neural Networks for Generic Structural Feature Extraction
• Achieving Better Kinship Recognition Through Better Baseline
• Unsupervised High-level Feature Learning by Ensemble Projection for Semi-supervised Image Classification and Image Clustering
• Progressive Feature Alignment for Unsupervised Domain Adaptation
• ConFoc: Content-Focus Protection Against Trojan Attacks on Neural Networks
• Cross Modal Distillation for Supervision Transfer
• Multiple instance learning on deep features for weakly supervised object detection with extreme domain shifts
• Factors of Transferability for a Generic ConvNet Representation
• Generic Deep Networks with Wavelet Scattering
• Analysing Deep Reinforcement Learning Agents Trained with Domain Randomisation
• GOGGLES: Automatic Image Labeling with Affinity Coding
• Learning to Generate Chairs, Tables and Cars with Convolutional Networks
• Cluster Alignment with a Teacher for Unsupervised Domain Adaptation
• Adversarial Reprogramming of Neural Networks
• Towards Practical Multi-Object Manipulation using Relational Reinforcement Learning
• Efficient Yet Deep Convolutional Neural Networks for Semantic Segmentation
• Scatter Component Analysis: A Unified Framework for Domain Adaptation and Domain Generalization
• Transferring GANs: generating images from limited data
• Brain-mediated Transfer Learning of Convolutional Neural Networks
• Transferable Calibration with Lower Bias and Variance in Domain Adaptation
• Selective Unsupervised Feature Learning with Convolutional Neural Network (S-CNN)
• Recognizing American Sign Language Manual Signs from RGB-D Videos
• Certified Robustness to Label-Flipping Attacks via Randomized Smoothing
• Mind the Nuisance: Gaussian Process Classification using Privileged Noise
• Landslide Detection in Real-Time Social Media Image Streams
• BREEDS: Benchmarks for Subpopulation Shift
• Few-Shot Image Recognition by Predicting Parameters from Activations
• On Last-Layer Algorithms for Classification: Decoupling Representation from Uncertainty Estimation
• An analysis of the transfer learning of convolutional neural networks for artistic images
• From handcrafted to deep local features
• Fast ConvNets Using Group-wise Brain Damage
• Image Credibility Analysis with Effective Domain Transferred Deep Networks
• Feature Evaluation of Deep Convolutional Neural Networks for Object Recognition and Detection
• Online Learning to Sample
• Robust Visual Tracking via Convolutional Networks
• Low-resource Deep Entity Resolution with Transfer and Active Learning
• Deep Attention-guided Hashing
• Semantic Change Detection with Hypermaps
• Bridging the Theoretical Bound and Deep Algorithms for Open Set Domain Adaptation
• The Effectiveness of Instance Normalization: a Strong Baseline for Single Image Dehazing
• Refining Language Models with Compositional Explanations
• Image Aesthetics Assessment Using Composite Features from off-the-Shelf Deep Models
• Transfer Learning for Estimating Causal Effects using Neural Networks
• Characterizing and Avoiding Negative Transfer
• Zero-shot Domain Adaptation without Domain Semantic Descriptors
• Temporal Attentive Alignment for Large-Scale Video Domain Adaptation
• Transfer Learning for Sequence Labeling Using Source Model and Target Data
• Embedding Label Structures for Fine-Grained Feature Representation
• DAG-Recurrent Neural Networks For Scene Labeling
• Transitive Hashing Network for Heterogeneous Multimedia Retrieval
• Seeing the Big Picture: Deep Embedding with Contextual Evidences
• Correlation Alignment by Riemannian Metric for Domain Adaptation
• Prototypical Cross-domain Self-supervised Learning for Few-shot Unsupervised Domain Adaptation
• Indoor Space Recognition using Deep Convolutional Neural Network: A Case Study at MIT Campus
• Fine-Grained Visual Categorization via Multi-stage Metric Learning
• Crime Mapping from Satellite Imagery via Deep Learning
• A Flexible Selection Scheme for Minimum-Effort Transfer Learning
• Constructing Deep Neural Networks by Bayesian Network Structure Learning
• Understanding Locally Competitive Networks
• Active Reward Learning for Co-Robotic Vision Based Exploration in Bandwidth Limited Environments
• Online Class-Incremental Continual Learning with Adversarial Shapley Value
• Time Matters: Multi-scale Temporalization of Social Media Popularity
• FCSS: Fully Convolutional Self-Similarity for Dense Semantic Correspondence
• Exploring the Design Space of Deep Convolutional Neural Networks at Large Scale
• Towards Practical Visual Search Engine within Elasticsearch
• On the energy landscape of deep networks
• A Secure Mobile Authentication Alternative to Biometrics
• Sample-level CNN Architectures for Music Auto-tagging Using Raw Waveforms
• Data-driven geophysics: from dictionary learning to deep learning
• Digging Deep into the layers of CNNs: In Search of How CNNs Achieve View Invariance
• A Survey of Visual Analytics Techniques for Machine Learning
• Towards Understanding the Transferability of Deep Representations
• Continual Learning in Neural Networks
• Online and Offline Handwritten Chinese Character Recognition: A Comprehensive Study and New Benchmark
• Modified Distribution Alignment for Domain Adaptation with Pre-trained Inception ResNet
• Outlining the Design Space of Explainable Intelligent Systems for Medical Diagnosis
• A Deep Generative Deconvolutional Image Model
• Self-supervised Modal and View Invariant Feature Learning
• Object Level Deep Feature Pooling for Compact Image Representation
• Zero-Shot Domain Adaptation via Kernel Regression on the Grassmannian
• Situational Object Boundary Detection
• TAFE-Net: Task-Aware Feature Embeddings for Low Shot Learning
• Look-into-Object: Self-supervised Structure Modeling for Object Recognition
• CLDA: Contrastive Learning for Semi-Supervised Domain Adaptation
• Beyond Fine-Tuning: Transferring Behavior in Reinforcement Learning
• Unifying Unsupervised Domain Adaptation and Zero-Shot Visual Recognition
• Scalable Greedy Algorithms for Transfer Learning
• Discriminative and Geometry Aware Unsupervised Domain Adaptation
• Learning Structured Inference Neural Networks with Label Relations
• Generalization through Simulation: Integrating Simulated and Real Data into Deep Reinforcement Learning for Vision-Based Autonomous Flight
• Ranking and Tuning Pre-trained Models: A New Paradigm for Exploiting Model Hubs
• Towards Creating a Deployable Grasp Type Probability Estimator for a Prosthetic Hand
• A Real-time System for Detecting Landslide Reports on Social Media using Artificial Intelligence
• Convolutional Models for Joint Object Categorization and Pose Estimation
• Unsupervised Domain Adaptation via Structured Prediction Based Selective Pseudo-Labeling
• Machine learning based co-creative design framework
• Apple Leaf Disease Identification through Region-of-Interest-Aware Deep Convolutional Neural Network
• Adaptive Denoising via GainTuning
• FRODO: Free rejection of out-of-distribution samples: application to chest x-ray analysis
• Simple and Efficient Learning using Privileged Information
• Discriminative Label Consistent Domain Adaptation
• Cross-Domain Visual Recognition via Domain Adaptive Dictionary Learning
• Partial Optimal Transport with Applications on Positive-Unlabeled Learning
• Advances in Deep Learning for Hyperspectral Image Analysis--Addressing Challenges Arising in Practical Imaging Scenarios
• MiniMax Entropy Network: Learning Category-Invariant Features for Domain Adaptation
• Accelerating Deep Unsupervised Domain Adaptation with Transfer Channel Pruning
• STNReID : Deep Convolutional Networks with Pairwise Spatial Transformer Networks for Partial Person Re-identification
• It Takes Two to Tango: Mixup for Deep Metric Learning
• MemeSequencer: Sparse Matching for Embedding Image Macros
• t-SNE-CUDA: GPU-Accelerated t-SNE and its Applications to Modern Data
• SuperTML: Two-Dimensional Word Embedding for the Precognition on Structured Tabular Data
• Deep Structured-Output Regression Learning for Computational Color Constancy
• Real-world Object Recognition with Off-the-shelf Deep Conv Nets: How Many Objects can iCub Learn?
• SA-CNN: Dynamic Scene Classification using Convolutional Neural Networks
• Changing Fashion Cultures
• DART: Domain-Adversarial Residual-Transfer Networks for Unsupervised Cross-Domain Image Classification
• What Do Deep Nets Learn? Class-wise Patterns Revealed in the Input Space
• Boosting Few-Shot Learning With Adaptive Margin Loss
• Contrastively Smoothed Class Alignment for Unsupervised Domain Adaptation
• Deep Learning for Surface Material Classification Using Haptic And Visual Information
• Semi-Supervised Co-Analysis of 3D Shape Styles from Projected Lines
• Robust Deep Sensing Through Transfer Learning in Cognitive Radio
• Object-Centric Representation Learning from Unlabeled Videos
• Adversarial Cross-Domain Action Recognition with Co-Attention
• Accelerating Deep Neural Networks with Spatial Bottleneck Modules
• Learning Fair and Transferable Representations
• Multi-scale Orderless Pooling of Deep Convolutional Activation Features
• Learning A Task-Specific Deep Architecture For Clustering
• A Spike Learning System for Event-driven Object Recognition
• Me, Myself and My Killfie: Characterizing and Preventing Selfie Deaths
• An Out-of-the-box Full-network Embedding for Convolutional Neural Networks
• Few-Shot Representation Learning for Out-Of-Vocabulary Words
• Deep causal representation learning for unsupervised domain adaptation
• Semantics for Global and Local Interpretation of Deep Neural Networks
• XMixup: Efficient Transfer Learning with Auxiliary Samples by Cross-domain Mixup
• Agent-Centric Representations for Multi-Agent Reinforcement Learning
• Data Augmentation with norm-VAE for Unsupervised Domain Adaptation
• Detecting Individual Decision-Making Style: Exploring Behavioral Stylometry in Chess
• CITISEN: A Deep Learning-Based Speech Signal-Processing Mobile Application
• Interactive Knowledge Distillation
• MineGAN: effective knowledge transfer from GANs to target domains with few images
• Deep Learning in Information Security
• SOSELETO: A Unified Approach to Transfer Learning and Training with Noisy Labels
• Scene Graph Parsing as Dependency Parsing
• Unsupervised Learning of Visual Representations by Solving Jigsaw Puzzles
• What is the right way to represent document images?
• Window-Object Relationship Guided Representation Learning for Generic Object Detections
• Pooling the Convolutional Layers in Deep ConvNets for Action Recognition
• Active Perceptual Similarity Modeling with Auxiliary Information
• Detector Discovery in the Wild: Joint Multiple Instance and Representation Learning
• Learning Smooth Representation for Unsupervised Domain Adaptation
• Attention-Privileged Reinforcement Learning
• Domain Adaptation by Mixture of Alignments of Second- or Higher-Order Scatter Tensors
• Towards Predicting the Likeability of Fashion Images
• CAGFuzz: Coverage-Guided Adversarial Generative Fuzzing Testing of Deep Learning Systems
• When Naïve Bayes Nearest Neighbours Meet Convolutional Neural Networks
• A contribution to Optimal Transport on incomparable spaces
• DW-GAN: A Discrete Wavelet Transform GAN for NonHomogeneous Dehazing
• Differentially Private Continual Learning
• Tag Prediction at Flickr: a View from the Darkroom
• Deep driven fMRI decoding of visual categories
• OoDAnalyzer: Interactive Analysis of Out-of-Distribution Samples
• End-to-end Image Captioning Exploits Multimodal Distributional Similarity
• Robustness via Deep Low-Rank Representations
• Unconstrained Face Verification using Deep CNN Features
• Context-Aware Multipath Networks
• Visual Language Modeling on CNN Image Representations
• Statistical Optimal Transport posed as Learning Kernel Embedding
• Learning From Less Data: Diversified Subset Selection and Active Learning in Image Classification Tasks
• Self-supervised Feature Learning by Cross-modality and Cross-view Correspondences
• Between-Domain Instance Transition Via the Process of Gibbs Sampling in RBM
• An Extended Framework for Marginalized Domain Adaptation
• Uncertainty Quantification for Deep Context-Aware Mobile Activity Recognition and Unknown Context Discovery
• Category-Based Deep CCA for Fine-Grained Venue Discovery from Multimodal Data
• Domain-Specific Suppression for Adaptive Object Detection
• GAN Cocktail: mixing GANs without dataset access
• DISA at ImageCLEF 2014 Revised: Search-based Image Annotation with DeCAF Features
• Adaptive Selection of Deep Learning Models on Embedded Systems
• Tailoring: encoding inductive biases by optimizing unsupervised objectives at prediction time
• Unsupervised Adversarial Graph Alignment with Graph Embedding
• Learning Effective Loss Functions Efficiently
• Group-wise Deep Co-saliency Detection
• CytoImageNet: A large-scale pretraining dataset for bioimage transfer learning
• BiPC: Bidirectional Probability Calibration for Unsupervised Domain Adaption
• Strength from Weakness: Fast Learning Using Weak Supervision
• Full-Network Embedding in a Multimodal Embedding Pipeline
• Happy Travelers Take Big Pictures: A Psychological Study with Machine Learning and Big Data
• EndoNet: A Deep Architecture for Recognition Tasks on Laparoscopic Videos
• Measuring and Understanding Sensory Representations within Deep Networks Using a Numerical Optimization Framework
• Part-Stacked CNN for Fine-Grained Visual Categorization
• Cross-Modal Similarity Learning : A Low Rank Bilinear Formulation
• Asymmetric Deep Semantic Quantization for Image Retrieval
• Discriminative Cross-Domain Feature Learning for Partial Domain Adaptation
• A Deeper Look at Dataset Bias
• Tell and Predict: Kernel Classifier Prediction for Unseen Visual Classes from Unstructured Text Descriptions
• Value of Temporal Dynamics Information in Driving Scene Segmentation
• HWNet v2: An Efficient Word Image Representation for Handwritten Documents
• Lightweight Unsupervised Domain Adaptation by Convolutional Filter Reconstruction
• Cross-Entropy Adversarial View Adaptation for Person Re-identification
• Selective Transfer with Reinforced Transfer Network for Partial Domain Adaptation
• Detecting retail products in situ using CNN without human effort labeling
• HoneyFaces: Increasing the Security and Privacy of Authentication Using Synthetic Facial Images
• Learning Rich Features from RGB-D Images for Object Detection and Segmentation
• HyperSTAR: Task-Aware Hyperparameters for Deep Networks
• Sample-based Regularization: A Transfer Learning Strategy Toward Better Generalization
• Predicting the Future Behavior of a Time-Varying Probability Distribution
• SORT: Second-Order Response Transform for Visual Recognition
• Analysing domain shift factors between videos and images for object detection
• A Combinatorial Perspective on Transfer Learning
• Mind the Gap: Subspace based Hierarchical Domain Adaptation
• Probability Weighted Compact Feature for Domain Adaptive Retrieval
• Unsupervised Transfer Learning for Spoken Language Understanding in Intelligent Agents
• Deeper, Broader and Artier Domain Generalization
• Distributed Deep Learning in Open Collaborations
• Domain Adaptation and Image Classification via Deep Conditional Adaptation Network
• Adversarial Training Helps Transfer Learning via Better Representations
• Building Graph Representations of Deep Vector Embeddings
• Deep Priority Hashing
• Deep Attributes from Context-Aware Regional Neural Codes
• Feature Stylization and Domain-aware Contrastive Learning for Domain Generalization
• CNN-based Action Recognition and Supervised Domain Adaptation on 3D Body Skeletons via Kernel Feature Maps
• On Machine Learning and Structure for Mobile Robots
• Part-based R-CNNs for Fine-grained Category Detection
• Learning from Videos with Deep Convolutional LSTM Networks
• Meta-learning with negative learning rates
• OTCE: A Transferability Metric for Cross-Domain Cross-Task Representations
• Matching Handwritten Document Images
• Motorcycle detection and classification in urban Scenarios using a model based on Faster R-CNN
• Cross-convolutional-layer Pooling for Image Recognition
• A Visual Embedding for the Unsupervised Extraction of Abstract Semantics
• A Deep Multi-task Learning Approach to Skin Lesion Classification
• Context Augmentation for Convolutional Neural Networks
• Locality Preserving Joint Transfer for Domain Adaptation
• The Treasure beneath Convolutional Layers: Cross-convolutional-layer Pooling for Image Classification
• Real-Time Freespace Segmentation on Autonomous Robots for Detection of Obstacles and Drop-Offs
• Large Scale Fine-Grained Categorization and Domain-Specific Transfer Learning
• Deep Triplet Quantization
• Heterogeneous Domain Adaptation via Soft Transfer Network
• Friction from Reflectance: Deep Reflectance Codes for Predicting Physical Surface Properties from One-Shot In-Field Reflectance
• Joint cross-domain classification and subspace learning for unsupervised adaptation
• Extrapolating Expected Accuracies for Large Multi-Class Problems
• Detection and Tracking of General Movable Objects in Large 3D Maps
• Semantic Hierarchical Priors for Intrinsic Image Decomposition
• MAFER: a Multi-resolution Approach to Facial Expression Recognition
• Facial Expression Recognition Research Based on Deep Learning
• New Perspectives on k-Support and Cluster Norms
• Structural Regularization
• Domain Conditioned Adaptation Network
• Attention Transfer Network for Aspect-level Sentiment Classification
• Regularity as Regularization: Smooth and Strongly Convex Brenier Potentials in Optimal Transport
• Visual Security Evaluation of Learnable Image Encryption Methods against Ciphertext-only Attacks
• CO-Optimal Transport
• On the Behavior of Convolutional Nets for Feature Extraction
• Learning a smooth kernel regularizer for convolutional neural networks
• Saliency-Aware Class-Agnostic Food Image Segmentation
• Adaptively-Accumulated Knowledge Transfer for Partial Domain Adaptation
• Patchy Image Structure Classification Using Multi-Orientation Region Transform
• Dual Memory Architectures for Fast Deep Learning of Stream Data via an Online-Incremental-Transfer Strategy
• Learned Indexes for Dynamic Workloads
• Pretraining Techniques for Sequence-to-Sequence Voice Conversion
• Predicting the Future: A Jointly Learnt Model for Action Anticipation
• Write a Classifier: Predicting Visual Classifiers from Unstructured Text
• A Systematic Comparison of Deep Learning Architectures in an Autonomous Vehicle
• Target Aware Network Adaptation for Efficient Representation Learning
• SMILE: Self-Distilled MIxup for Efficient Transfer LEarning
• Compression of Deep Neural Networks on the Fly
• Local Color Contrastive Descriptor for Image Classification
• Transfer Learning Using Classification Layer Features of CNN
• Learning Mid-Level Features and Modeling Neuron Selectivity for Image Classification
• Hyperparameter-free Continuous Learning for Domain Classification in Natural Language Understanding
• GAN Memory with No Forgetting
• From Anchor Generation to Distribution Alignment: Learning a Discriminative Embedding Space for Zero-Shot Recognition
• Material Classification using Neural Networks
• Learn Faster and Forget Slower via Fast and Stable Task Adaptation
• Multi-source Heterogeneous Domain Adaptation with Conditional Weighting Adversarial Network
• Reverse Transfer Learning: Can Word Embeddings Trained for Different NLP Tasks Improve Neural Language Models?
• Effective Domain Knowledge Transfer with Soft Fine-tuning
• A Primal-Dual Subgradient Approachfor Fair Meta Learning
• Visualizing Image Content to Explain Novel Image Discovery
• A feasibility study of deep neural networks for the recognition of banknotes regarding central bank requirements
• Global Semantic Description of Objects based on Prototype Theory
• Efficient Object Embedding for Spliced Image Retrieval
• Ranking Neural Checkpoints
• A Fully Spiking Hybrid Neural Network for Energy-Efficient Object Detection
• Pretrained Encoders are All You Need
• Introducing the structural bases of typicality effects in deep learning
• Developing efficient transfer learning strategies for robust scene recognition in mobile robotics using pre-trained convolutional neural networks
• Visual Domain Adaptation with Manifold Embedded Distribution Alignment
• Multi-source Few-shot Domain Adaptation
• A Simple Approach to Continual Learning by Transferring Skill Parameters
• Characterizing and Understanding the Generalization Error of Transfer Learning with Gibbs Algorithm
• Pre-training without Natural Images
• Drive Video Analysis for the Detection of Traffic Near-Miss Incidents
• Learning Invariant Representations across Domains and Tasks
• Online Convolutional Sparse Coding with Sample-Dependent Dictionary
• Group Based Deep Shared Feature Learning for Fine-grained Image Classification
• Towards Accurate and Robust Domain Adaptation under Noisy Environments
• Image Transformation Network for Privacy-Preserving Deep Neural Networks and Its Security Evaluation
• A Computational Approach to Relative Aesthetics
• Transfer Learning in 4D for Breast Cancer Diagnosis using Dynamic Contrast-Enhanced Magnetic Resonance Imaging
• PAC-Bayes Analysis of Sentence Representation
• Joint Concept Matching based Learning for Zero-Shot Recognition
• Adaptive Compression-based Lifelong Learning
• Make Your Bone Great Again : A study on Osteoporosis Classification
• Modeling Image Virality with Pairwise Spatial Transformer Networks
• Beat-Event Detection in Action Movie Franchises
• Efficient Inferencing of Compressed Deep Neural Networks
• Mining Mid-level Visual Patterns with Deep CNN Activations
• Learning to Select Pre-Trained Deep Representations with Bayesian Evidence Framework
• Personality, Culture, and System Factors - Impact on Affective Response to Multimedia
• Visual Themes and Sentiment on Social Networks To Aid First Responders During Crisis Events
• Robust Visual Knowledge Transfer via EDA
• An Overview on Data Representation Learning: From Traditional Feature Learning to Recent Deep Learning
• Compressive Sensing via Convolutional Factor Analysis
• Towards Unsupervised Weed Scouting for Agricultural Robotics
• Hardware-Driven Nonlinear Activation for Stochastic Computing Based Deep Convolutional Neural Networks
• Coupled Depth Learning
• Persistent Evidence of Local Image Properties in Generic ConvNets
• Accelerated kernel discriminant analysis
• CRL: Class Representative Learning for Image Classification
• Domain2Vec: Domain Embedding for Unsupervised Domain Adaptation
• Leveraging Visual Question Answering for Image-Caption Ranking
• Adapting Convolutional Neural Networks for Geographical Domain Shift
• Structured Sparse Convolutional Autoencoder
• Towards Robust Pattern Recognition: A Review
• Unsupervised Domain Expansion from Multiple Sources
• SERIL: Noise Adaptive Speech Enhancement using Regularization-based Incremental Learning
• Unsupervised Deep Domain Adaptation for Pedestrian Detection
• Unsupervised Domain Adaptation: A Multi-task Learning-based Method
• Compositional Model based Fisher Vector Coding for Image Classification
• Learning Optimal Linear Regularizers
• RCoNet: Deformable Mutual Information Maximization and High-order Uncertainty-aware Learning for Robust COVID-19 Detection
• Manifold Criterion Guided Transfer Learning via Intermediate Domain Generation
• Feature-Level Domain Adaptation
• Amortized Object and Scene Perception for Long-term Robot Manipulation
• Discriminative Noise Robust Sparse Orthogonal Label Regression-based Domain Adaptation
• Towards Recognizing New Semantic Concepts in New Visual Domains
• Meta-learning Transferable Representations with a Single Target Domain
• A Two-Layer Local Constrained Sparse Coding Method for Fine-Grained Visual Categorization
• Self-Adaptive Partial Domain Adaptation
• Growing a Brain: Fine-Tuning by Increasing Model Capacity
• Understand Scene Categories by Objects: A Semantic Regularized Scene Classifier Using Convolutional Neural Networks
• GAPLE: Generalizable Approaching Policy LEarning for Robotic Object Searching in Indoor Environment
• Unsupervised Cross-Domain Recognition by Identifying Compact Joint Subspaces
• Domain Generalization under Conditional and Label Shifts via Variational Bayesian Inference
• Few-shot Learning for Unsupervised Feature Selection
• Why Do Better Loss Functions Lead to Less Transferable Features?
• Exemplar Based Deep Discriminative and Shareable Feature Learning for Scene Image Classification
• PanRep: Graph neural networks for extracting universal node embeddings in heterogeneous graphs
• SCPNet: Spatial-Channel Parallelism Network for Joint Holistic and Partial Person Re-Identification
• One-Shot Item Search with Multimodal Data
• Multi-lingual agents through multi-headed neural networks
• An active search strategy for efficient object class detection
• Large e-retailer image dataset for visual search and product classification
• Fabric Surface Characterization: Assessment of Deep Learning-based Texture Representations Using a Challenging Dataset
• Practical Transferability Estimation for Image Classification Tasks
• Efficient Convolutional Neural Network with Binary Quantization Layer
• Convolutional Channel Features
• Hashing in the Zero Shot Framework with Domain Adaptation
• Improve CAM with Auto-adapted Segmentation and Co-supervised Augmentation
• Representation as a Service
• (DE)^2 CO: Deep Depth Colorization
• Objects as context for detecting their semantic parts
• Prior Knowledge about Attributes: Learning a More Effective Potential Space for Zero-Shot Recognition
• Classification of Radio Signals Using Truncated Gaussian Discriminant Analysis of Convolutional Neural Network-Derived Features
• Recovering 6D Object Pose: A Review and Multi-modal Analysis
• Bi-Directional Generation for Unsupervised Domain Adaptation
• Defending Against Adversarial Attacks Using Random Forests
• Weakly-Supervised Spatial Context Networks
• A Machine Learning Framework for Automatic Prediction of Human Semen Motility
• Domain Adaptation by Maximizing Population Correlation with Neural Architecture Search
• Open Set Domain Adaptation for Image and Action Recognition
• Deep Affordance-grounded Sensorimotor Object Recognition
• Reconstructing A Large Scale 3D Face Dataset for Deep 3D Face Identification
• Toward Optimal Run Racing: Application to Deep Learning Calibration
• Fine-Grained Categorization via CNN-Based Automatic Extraction and Integration of Object-Level and Part-Level Features
• An Analysis of Human-centered Geolocation
• Local Higher-Order Statistics (LHS) describing images with statistics of local non-binarized pixel patterns
• VisGraphNet: a complex network interpretation of convolutional neural features
• Towards Accurate Knowledge Transfer via Target-awareness Representation Disentanglement
• cvpaper.challenge in 2016: Futuristic Computer Vision through 1,600 Papers Survey
• Identifying and Exploiting Structures for Reliable Deep Learning
• Self-Tuning for Data-Efficient Deep Learning
• Matching the Clinical Reality: Accurate OCT-Based Diagnosis From Few Labels
• Hard Class Rectification for Domain Adaptation
• Universality of Gradient Descent Neural Network Training
• Conformal retrofitting via Riemannian manifolds: distilling task-specific graphs into pretrained embeddings
• Cross-domain error minimization for unsupervised domain adaptation
• Using Cross-Model EgoSupervision to Learn Cooperative Basketball Intention
• Deep Adversarial Domain Adaptation Based on Multi-layer Joint Kernelized Distance
• Semantic Graph for Zero-Shot Learning
• Research Frontiers in Transfer Learning -- a systematic and bibliometric review
• Understanding Human Judgments of Causality
• Theoretical and Experimental Analysis on the Generalizability of Distribution Regression Network
• Image Annotation Incorporating Low-Rankness, Tag and Visual Correlation and Inhomogeneous Errors
• Transformative Machine Learning
• Multiple Subspace Alignment Improves Domain Adaptation
• Toward Multimodal Modeling of Emotional Expressiveness
• Learning Rich Representations For Structured Visual Prediction Tasks
• Can We Teach Computers to Understand Art? Domain Adaptation for Enhancing Deep Networks Capacity to De-Abstract Art
• Adaptive Transfer Learning: a simple but effective transfer learning
• Marine Animal Classification with Correntropy Loss Based Multi-view Learning
• Understanding the Dynamics of DNNs Using Graph Modularity
• Engineering Deep Representations for Modeling Aesthetic Perception
• TAGLETS: A System for Automatic Semi-Supervised Learning with Auxiliary Data
• RSAC: Regularized Subspace Approximation Classifier for Lightweight Continuous Learning
• Multi-component Image Translation for Deep Domain Generalization
• Learning Pixel Representations for Generic Segmentation
• Local Area Transform for Cross-Modality Correspondence Matching and Deep Scene Recognition
• Bilinear Supervised Hashing Based on 2D Image Features
• DASC: Robust Dense Descriptor for Multi-modal and Multi-spectral Correspondence Estimation
• Low-Cost Transfer Learning of Face Tasks
• "Who is Driving around Me?" Unique Vehicle Instance Classification using Deep Neural Features
• Renofeation: A Simple Transfer Learning Method for Improved Adversarial Robustness
• Unsupervised Domain Adaptation in the Wild: Dealing with Asymmetric Label Sets
• Transfer Learning Based on AdaBoost for Feature Selection from Multiple ConvNet Layer Features
• A Simple Domain Shifting Networkfor Generating Low Quality Images
• Geometric Neural Phrase Pooling: Modeling the Spatial Co-occurrence of Neurons
• Learning Joint Representations of Videos and Sentences with Web Image Search
• Adversarial Transfer Learning for Cross-domain Visual Recognition
• Location Recognition Over Large Time Lags
• Unsupervised Representation Learning with Laplacian Pyramid Auto-encoders
• A novel learning-based frame pooling method for Event Detection
• Action Segmentation with Mixed Temporal Domain Adaptation
• Weakly Supervised Object Localization and Detection: A Survey
• Learning to see across Domains and Modalities
• Combining Diverse Feature Priors
• Cross-domain Image Retrieval with a Dual Attribute-aware Ranking Network
• DAP: Detection-Aware Pre-training with Weak Supervision
• Supervised Initialization of LSTM Networks for Fundamental Frequency Detection in Noisy Speech Signals
• Towards Making Deep Transfer Learning Never Hurt
• Subspace Clustering Based Tag Sharing for Inductive Tag Matrix Refinement with Complex Errors
• Unlabeled Data Deployment for Classification of Diabetic Retinopathy Images Using Knowledge Transfer
• Identifying Reliable Annotations for Large Scale Image Segmentation
• Dual Pattern Learning Networks by Empirical Dual Prediction Risk Minimization
• Multi-Scale Spatially-Asymmetric Recalibration for Image Classification
• Explicit homography estimation improves contrastive self-supervised learning
• Subset Feature Learning for Fine-Grained Category Classification
• LOTS about Attacking Deep Features
• Joint Learning of Discriminative Low-dimensional Image Representations Based on Dictionary Learning and Two-layer Orthogonal Projections
• DWMD: Dimensional Weighted Orderwise Moment Discrepancy for Domain-specific Hidden Representation Matching
• Improving Voice Separation by Incorporating End-to-end Speech Recognition
• Latent Hinge-Minimax Risk Minimization for Inference from a Small Number of Training Samples
• Joint calibration of Ensemble of Exemplar SVMs
• Deep Virtual Networks for Memory Efficient Inference of Multiple Tasks
• Filtered Manifold Alignment
• Transfer Learning with Sparse Associative Memories
• Recent Advancements in Self-Supervised Paradigms for Visual Feature Representation
• Hierarchically Robust Representation Learning
• Meta-Learning to Improve Pre-Training