Domain Stylization: A Strong, Simple Baseline for Synthetic to Real Image Domain Adaptation

Abstract

Deep neural networks have largely failed to effectively utilize synthetic data when applied to real images due to the covariate shift problem. In this paper, we show that by applying a straightforward modification to an existing photorealistic style transfer algorithm, we achieve state-of-the-art synthetic-to-real domain adaptation results. We conduct extensive experimental validations on four synthetic-to-real tasks for semantic segmentation and object detection, and show that our approach exceeds the performance of any current state-of-the-art GAN-based image translation approach as measured by segmentation and object detection metrics. Furthermore we offer a distance based analysis of our method which shows a dramatic reduction in Frechet Inception distance between the source and target domains, offering a quantitative metric that demonstrates the effectiveness of our algorithm in bridging the synthetic-to-real gap.

References in corpus (14)

• Rethinking Atrous Convolution for Semantic Image Segmentation
• GANs Trained by a Two Time-Scale Update Rule Converge to a Local Nash Equilibrium
• Deep Domain Confusion: Maximizing for Domain Invariance
• Unpaired Image-to-Image Translation using Cycle-Consistent Adversarial Networks
• The Cityscapes Dataset for Semantic Urban Scene Understanding
• FCNs in the Wild: Pixel-level Adversarial and Constraint-based Adaptation
• Unsupervised Image-to-Image Translation Networks
• Unsupervised Cross-Domain Image Generation
• Multimodal Unsupervised Image-to-Image Translation
• Building Generalizable Agents with a Realistic and Rich 3D Environment
• Domain Randomization for Transferring Deep Neural Networks from Simulation to the Real World
• Dilated Residual Networks
• Playing for Data: Ground Truth from Computer Games
• A Closed-form Solution to Photorealistic Image Stylization

Cited by in corpus (23)

• Self-supervised Domain Adaptation for Computer Vision Tasks
• Learning Linear Transformations for Fast Arbitrary Style Transfer
• Adversarial Style Mining for One-Shot Unsupervised Domain Adaptation
• Unsupervised Domain Adaptation for Mobile Semantic Segmentation based on Cycle Consistency and Feature Alignment
• Self-Ensembling with GAN-based Data Augmentation for Domain Adaptation in Semantic Segmentation
• Unsupervised Domain Adaptation in Semantic Segmentation: a Review
• Dynamic Instance Normalization for Arbitrary Style Transfer
• Condition-Invariant Semantic Segmentation
• Cut-and-Paste Dataset Generation for Balancing Domain Gaps in Object Instance Detection
• A system for generating complex physically accurate sensor images for automotive applications
• MAT: A simple yet strong baseline for identifying self-admitted technical debt
• Adversarial Appearance Learning in Augmented Cityscapes for Pedestrian Recognition in Autonomous Driving
• Domain Adaptation for Image Dehazing
• Towards Domain Invariant Single Image Dehazing
• Beyond Domain Adaptation: Unseen Domain Encapsulation via Universal Non-volume Preserving Models
• Soft Prototyping Camera Designs for Car Detection Based on a Convolutional Neural Network
• Transferring and Regularizing Prediction for Semantic Segmentation
• DLOW: Domain Flow for Adaptation and Generalization
• Synthetic training data generation for deep learning based quality inspection
• CrDoCo: Pixel-level Domain Transfer with Cross-Domain Consistency
• Multiple Style Transfer via Variational AutoEncoder
• Domain-robust VQA with diverse datasets and methods but no target labels
• Domain Adaptation for Rare Classes Augmented with Synthetic Samples