Uncertainty-Guided Progressive GANs for Medical Image Translation

Upadhyay, Uddeshya; Chen, Yanbei; Hepp, Tobias; Gatidis, Sergios; Akata, Zeynep

doi:10.1007/978-3-030-87199-4_58

Cited by 18 publications

(13 citation statements)

References 24 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The deep recurrent attentive writer (DRAW) [36] predicts the data content part‐by‐part through an attention mechanism, and the LAPGAN [30] learns in a coarse‐to‐fine manner. The UP‐GAN [12] uses uncertainty to guide the inference process. Our approach improves upon the state‐of‐the‐art models in terms of multiple measures at the cost of a small amount of complexity increment for the Shining3D tooth dataset.…”

Section: Resultsmentioning

confidence: 99%

“…The relativistic GAN [11] addresses the observation that generator training increases the probability that fake data are real, which decreases the probability that real data are actually real. The uncertainty-guided progressive GAN (UP-GAN) [12] was developed to explore the uncertainty of inference with respect to guiding the image generation process; however, the results lacked interpretability.…”

Section: Generative Adversarial Networkmentioning

confidence: 99%

See 1 more Smart Citation

A tooth surface design method combining semantic guidance, confidence, and structural coherence

Wang

Tian

Liu

et al. 2022

IET Computer Vision

View full text Add to dashboard Cite

Research on tooth surface design based on deep neural networks has recently achieved progress in terms of both accuracy and execution efficiency. However, unrealistic outputs are still a challenging issue, partially because of (1) the lack of semantic guidance, (2) the inability to discover and rectify false results, and (3) the lack of exploration of structural coherence in intermediate layers. In this paper, we present an approach to predict depth images for designed teeth based on a conditional generative adversarial network (CGAN) by incorporating semantic guidance. Moreover, the uncertainty of semantic inference is employed to improve the model outputs, and a structural coherence loss is proposed for adversarial learning to enhance the discrimination capability of the network in intermediate layers. We evaluate the performance of our approach with the Shining3D tooth dataset. The experimental results show that our method produces better results than the other available approaches in terms of accuracy.This is an open access article under the terms of the Creative Commons Attribution-NonCommercial License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited and is not used for commercial purposes.

show abstract

Section: Resultsmentioning

confidence: 99%

Section: Generative Adversarial Networkmentioning

confidence: 99%

A tooth surface design method combining semantic guidance, confidence, and structural coherence

Wang

Tian

Liu

et al. 2022

IET Computer Vision

View full text Add to dashboard Cite

show abstract

“…UP-GAN We adapted the UP-GAN network from [18] to run on unpaired dataset as shown in [17]. UP-GAN uses an uncertainty guided loss along the standard cycle loss during training.…”

Section: Network Architectures and Implementation Detailsmentioning

confidence: 99%

“…If the generated images are of so low quality, that the segmentor fails completely (DICE < 0.5) the confidence value does not correlate with the DICE score. We compare our method to the existing way of estimating aleatoric uncertainty, described in UP-GAN [18]. The quality of generated images is measured with FID scores and the correlation between the DICE coefficient and the mean of the estimated aleatoric uncertainty values as defined in [17].…”

Section: Can We Use the Noise Injections To Improve The Quality Of A ...mentioning

confidence: 99%

“…The issue has a considerable impact on the medical field, where generated images with fabricated features have no clinical value. Recently, Upadhyay et al [18] tackled this problem by predicting not only the output images but also the corresponding aleatoric uncertainty and then using it to guide the GAN to improve the final output. Their method requires changes in the optimization process (additional loss terms) and network architecture (additional output).…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

What can we learn about a generated image corrupting its latent representation?

Tomczak¹,

Gupta²,

Ilić³

et al. 2022

Preprint

View full text Add to dashboard Cite

Generative adversarial networks (GANs) offer an effective solution to the image-to-image translation problem, thereby allowing for new possibilities in medical imaging. They can translate images from one imaging modality to another at a low cost. For unpaired datasets, they rely mostly on cycle loss. Despite its effectiveness in learning the underlying data distribution, it can lead to a discrepancy between input and output data. The purpose of this work is to investigate the hypothesis that we can predict image quality based on its latent representation in the GANs bottleneck. We achieve this by corrupting the latent representation with noise and generating multiple outputs. The degree of differences between them is interpreted as the strength of the representation: the more robust the latent representation, the fewer changes in the output image the corruption causes. Our results demonstrate that our proposed method has the ability to i) predict uncertain parts of synthesized images, and ii) identify samples that may not be reliable for downstream tasks, e.g., liver segmentation task.

show abstract