Robust Semantic Segmentation with Superpixel-Mix
arXiv:2108.00968
Abstract
Along with predictive performance and runtime speed, reliability is a key requirement for real-world semantic segmentation. Reliability encompasses robustness, predictive uncertainty and reduced bias. To improve reliability, we introduce Superpixel-mix, a new superpixel-based data augmentation method with teacher-student consistency training. Unlike other mixing-based augmentation techniques, mixing superpixels between images is aware of object boundaries, while yielding consistent gains in segmentation accuracy. Our proposed technique achieves state-of-the-art results in semi-supervised semantic segmentation on the Cityscapes dataset. Moreover, Superpixel-mix improves the reliability of semantic segmentation by reducing network uncertainty and bias, as confirmed by competitive results under strong distributions shift (adverse weather, image corruptions) and when facing out-of-distribution data.
arXiv commentAccepted to BMVC2021
References in corpus (28)
- PyTorch: An Imperative Style, High-Performance Deep Learning Library
- Encoder-Decoder with Atrous Separable Convolution for Semantic Image Segmentation
- What Uncertainties Do We Need in Bayesian Deep Learning for Computer Vision?
- Improved Regularization of Convolutional Neural Networks with Cutout
- Shortcut Learning in Deep Neural Networks
- A Baseline for Detecting Misclassified and Out-of-Distribution Examples in Neural Networks
- Mean teachers are better role models: Weight-averaged consistency targets improve semi-supervised deep learning results
- Semantic Foggy Scene Understanding with Synthetic Data
- DeepXplore: Automated Whitebox Testing of Deep Learning Systems
- The Cityscapes Dataset for Semantic Urban Scene Understanding
- High-Resolution Representations for Labeling Pixels and Regions
- ImageNet-trained CNNs are biased towards texture; increasing shape bias improves accuracy and robustness
- CutMix: Regularization Strategy to Train Strong Classifiers with Localizable Features
- Evidential Deep Learning to Quantify Classification Uncertainty
- Robust Physical-World Attacks on Deep Learning Models
- Manifold Mixup: Better Representations by Interpolating Hidden States
- Deep Ensembles: A Loss Landscape Perspective
- Why M Heads are Better than One: Training a Diverse Ensemble of Deep Networks
- ReMixMatch: Semi-Supervised Learning with Distribution Alignment and Augmentation Anchoring
- Examining the Impact of Blur on Recognition by Convolutional Networks
- Benchmarking Robustness in Object Detection: Autonomous Driving when Winter is Coming
- BatchEnsemble: An Alternative Approach to Efficient Ensemble and Lifelong Learning
- Pitfalls of In-Domain Uncertainty Estimation and Ensembling in Deep Learning
- Puzzle Mix: Exploiting Saliency and Local Statistics for Optimal Mixup
- DFANet: Deep Feature Aggregation for Real-Time Semantic Segmentation
- Robustness to Spurious Correlations via Human Annotations
- On Last-Layer Algorithms for Classification: Decoupling Representation from Uncertainty Estimation
- Encoding the latent posterior of Bayesian Neural Networks for uncertainty quantification