AI Paper Atlas

Research map / Label-efficient and robust learning / Domain adaptation and generalization

DeceptionNet: Network-Driven Domain Randomization

Sergey Zakharov, Wadim Kehl, Slobodan Ilic

ICCV 2019 · Label-efficient and robust learning › Domain adaptation and generalization · approach: Geometric alignment and synthetic-to-real adaptation · 82 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
guide augmentation with the adversarial deception network instead of unguided random perturbationDeceptionNet · MNIST to MNIST-Maccuracy: 83.1 -> 90.4Improved
guide augmentation with the deception network instead of unguided random perturbationDeceptionNet · Synthetic to Real Cropped LineMODaccuracy: 53.1 -> 95.8Improved
guide augmentation with the deception network instead of unguided random perturbationDeceptionNet · SYNTHIA to CityscapesmIoU: 22.3 -> 27.5Improved
add background module (ablation)DeceptionNet · MNIST to MNIST-Maccuracy: 56.6 -> 82.4Improved
add noise module to background module (ablation)DeceptionNet · MNIST to MNIST-Maccuracy: 82.4 -> 86.5Improved
add distortion module to background and noise (ablation)DeceptionNet · MNIST to MNIST-Maccuracy: 86.5 -> 90.4Improved
add noise module to background module (ablation)DeceptionNet · Synthetic to Real Cropped LineMODaccuracy and angle error: 74.8 -> 77.6 accuracy, 50.4 -> 52.8 angle errorMixed
add distortion module to background and noise (ablation)DeceptionNet · Synthetic to Real Cropped LineMODaccuracy and angle error: 77.6 -> 78.7 accuracy, 52.8 -> 48.2 angle errorImproved
add light module (ablation)DeceptionNet · Synthetic to Real Cropped LineMODaccuracy and angle error: 78.7 -> 95.8 accuracy, 48.2 -> 51.9 angle errorMixed
use RGB-D input instead of RGB only (ablation)DeceptionNet · Synthetic to Real Cropped LineMODaccuracy: 84.8 -> 95.8Improved
use RGB-D input instead of depth only (ablation)DeceptionNet · Synthetic to Real Cropped LineMODaccuracy: 73.3 -> 95.8Improved
use RGB-D input instead of RGB only (ablation)DeceptionNet · Synthetic to Real Cropped LineMODmean angle error: 57.4 -> 51.9Improved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy