Road layout randomization on semantic labels produces synthetic training pairs that improve mIoU for rare road marking classes by over 12 percentage points in real-world urban deployment while retaining performance on other classes.
Domain Stylization: A Strong, Simple Baseline for Synthetic to Real Image Domain Adaptation
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
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.
years
2019 2verdicts
UNVERDICTED 2representative citing papers
Lightweight input adapters preprocess images to match ideal-condition training data for off-the-shelf CV models, enabling self-supervised incremental adaptation and reported gains in segmentation and localization on RobotCar and BDD datasets.
citing papers explorer
-
Generating All the Roads to Rome: Road Layout Randomization for Improved Road Marking Segmentation
Road layout randomization on semantic labels produces synthetic training pairs that improve mIoU for rare road marking classes by over 12 percentage points in real-world urban deployment while retaining performance on other classes.
-
Don't Worry About the Weather: Unsupervised Condition-Dependent Domain Adaptation
Lightweight input adapters preprocess images to match ideal-condition training data for off-the-shelf CV models, enabling self-supervised incremental adaptation and reported gains in segmentation and localization on RobotCar and BDD datasets.