Learning Representations for Automatic Colorization

Abstract

We develop a fully automatic image colorization system. Our approach leverages recent advances in deep networks, exploiting both low-level and semantic representations. As many scene elements naturally appear according to multimodal color distributions, we train our model to predict per-pixel color histograms. This intermediate output can be used to automatically generate a color image, or further manipulated prior to image formation. On both fully and partially automatic colorization tasks, we outperform existing methods. We also explore colorization as a vehicle for self-supervised visual representation learning.

ECCV 2016 (Project page: http://people.cs.uchicago.edu/~larsson/colorization/)

References in corpus (3)

• Caffe: Convolutional Architecture for Fast Feature Embedding
• Adversarial Feature Learning
• Adversarially Learned Inference

Cited by in corpus (52)

• Unsupervised Representation Learning by Predicting Image Rotations
• What makes ImageNet good for transfer learning?
• Deep Nearest Neighbor Anomaly Detection
• PixelNet: Representation of the pixels, by the pixels, and for the pixels
• Deep Koalarization: Image Colorization using CNNs and Inception-ResNet-v2
• Fake Colorized Image Detection
• Contrastive Learning with Adversarial Examples
• Exposure: A White-Box Photo Post-Processing Framework
• Colorization as a Proxy Task for Visual Understanding
• Less is More: Surgical Phase Recognition with Less Annotations through Self-Supervised Pre-training of CNN-LSTM Networks
• Point-BERT: Pre-training 3D Point Cloud Transformers with Masked Point Modeling
• Outline Colorization through Tandem Adversarial Networks
• Image Companding and Inverse Halftoning using Deep Convolutional Neural Networks
• 3D Shape Reconstruction from Sketches via Multi-view Convolutional Networks
• A Study of the Generalizability of Self-Supervised Representations
• Real-Time User-Guided Image Colorization with Learned Deep Priors
• Colorful Image Colorization
• Unsupervised Representation Learning by Sorting Sequences
• Scribbler: Controlling Deep Image Synthesis with Sketch and Color
• Deep Feature Consistent Deep Image Transformations: Downscaling, Decolorization and HDR Tone Mapping
• TextureGAN: Controlling Deep Image Synthesis with Texture Patches
• Transitive Invariance for Self-supervised Visual Representation Learning
• Deep Image Harmonization
• An inner-loop free solution to inverse problems using deep neural networks
• Ranking CGANs: Subjective Control over Semantic Image Attributes
• Split-Brain Autoencoders: Unsupervised Learning by Cross-Channel Prediction
• Exploring Instance Relations for Unsupervised Feature Embedding
• Unsupervised Deep Metric Learning via Auxiliary Rotation Loss
• Understand and Improve Contrastive Learning Methods for Visual Representation: A Review
• X-GANs: Image Reconstruction Made Easy for Extreme Cases
• Learning Diverse Image Colorization
• Self-Supervised Relative Depth Learning for Urban Scene Understanding
• On The State of Data In Computer Vision: Human Annotations Remain Indispensable for Developing Deep Learning Models
• Mix-and-Match Tuning for Self-Supervised Semantic Segmentation
• PixelNN: Example-based Image Synthesis
• InverseNet: Solving Inverse Problems with Splitting Networks
• Video Jigsaw: Unsupervised Learning of Spatiotemporal Context for Video Action Recognition
• Generalized Deep Image to Image Regression
• Play as You Like: Timbre-enhanced Multi-modal Music Style Transfer
• Fully Automatic Video Colorization with Self-Regularization and Diversity
• Awesome Typography: Statistics-Based Text Effects Transfer
• Sparse Factorization Layers for Neural Networks with Limited Supervision
• ColorUNet: A convolutional classification approach to colorization
• Towards Robust Pattern Recognition: A Review
• Modelling the Scene Dependent Imaging in Cameras with a Deep Neural Network
• Perceive Where to Focus: Learning Visibility-aware Part-level Features for Partial Person Re-identification
• Learning Rich Representations For Structured Visual Prediction Tasks
• Self-supervised Knowledge Distillation Using Singular Value Decomposition
• Deep Semantics-Aware Photo Adjustment
• Emerging Applications of Reversible Data Hiding
• Inner Space Preserving Generative Pose Machine
• Exploit Clues from Views: Self-Supervised and Regularized Learning for Multiview Object Recognition