Corpus ID: 236912480

Calibrated Adversarial Refinement for Stochastic Semantic Segmentation

  title={Calibrated Adversarial Refinement for Stochastic Semantic Segmentation},
  author={Elias Kassapis and G. Dikov and Deepak K. Gupta and C. Nugteren},
In semantic segmentation tasks, input images can often have more than one plausible interpretation, thus allowing for multiple valid labels. To capture such ambiguities, recent work has explored the use of probabilistic networks that can learn a distribution over predictions. However, these do not necessarily represent the empirical distribution accurately. In this work, we present a strategy for learning a calibrated predictive distribution over semantic maps, where the probability associated… Expand


I Bet You Are Wrong: Gambling Adversarial Networks for Structured Semantic Segmentation
This paper rethink adversarial training for semantic segmentation and proposes to reformulate the fake/real discrimination framework with a correct/incorrect training objective, replacing the discriminator with a "gambler" network that learns to spot and distribute its budget in areas where the predictions are clearly wrong. Expand
EL-GAN: Embedding Loss Driven Generative Adversarial Networks for Lane Detection
This work proposes EL-GAN: a GAN framework to mitigate the discussed problem using an embedding loss, and uses the TuSimple lane marking challenge to demonstrate that with this proposed framework it is viable to overcome the inherent anomalies of posing it as a semantic segmentation problem. Expand
Semantic Segmentation using Adversarial Networks
An adversarial training approach to train semantic segmentation models that can detect and correct higher-order inconsistencies between ground truth segmentation maps and the ones produced by the segmentation net. Expand
A Hierarchical Probabilistic U-Net for Modeling Multi-Scale Ambiguities
The Hierarchical Probabilistic U-Net is proposed, a segmentation network with a conditional variational auto-encoder (cVAE) that uses a hierarchical latent space decomposition that automatically separates independent factors across scales, an inductive bias that is deemed beneficial in structured output prediction tasks beyond segmentation. Expand
A Probabilistic U-Net for Segmentation of Ambiguous Images
A generative segmentation model based on a combination of a U-Net with a conditional variational autoencoder that is capable of efficiently producing an unlimited number of plausible hypotheses and reproduces the possible segmentation variants as well as the frequencies with which they occur significantly better than published approaches. Expand
Evaluating Bayesian Deep Learning Methods for Semantic Segmentation
This work modify DeepLab-v3+, one of the state-of-the-art deep neural networks, and create its Bayesian counterpart using MC dropout and Concrete dropout as inference techniques, and compares and test these two inference techniques on the well-known Cityscapes dataset using the suggested metrics. Expand
DeepLab: Semantic Image Segmentation with Deep Convolutional Nets, Atrous Convolution, and Fully Connected CRFs
This work addresses the task of semantic image segmentation with Deep Learning and proposes atrous spatial pyramid pooling (ASPP), which is proposed to robustly segment objects at multiple scales, and improves the localization of object boundaries by combining methods from DCNNs and probabilistic graphical models. Expand
CVAE-GAN: Fine-Grained Image Generation through Asymmetric Training
We present variational generative adversarial networks, a general learning framework that combines a variational auto-encoder with a generative adversarial network, for synthesizing images inExpand
Diversity-Sensitive Conditional Generative Adversarial Networks
It is shown that simple addition of the proposed regularization to existing models leads to surprisingly diverse generations, substantially outperforming the previous approaches for multi-modal conditional generation specifically designed in each individual task. Expand
Variational Approaches for Auto-Encoding Generative Adversarial Networks
This paper develops a principle upon which auto-encoders can be combined with generative adversarial networks by exploiting the hierarchical structure of the generative model, and describes a unified objective for optimization. Expand