Semi-Supervised Automatic Layer and Fluid Region Segmentation of Retinal Optical Coherence Tomography Images Using Adversarial Learning

Liu, Xiaoming; Fu, Tianyu; Pan, Zhifang; Liu, Dong; Hu, Wei; Li, Bo

doi:10.1109/icip.2018.8451071

Cited by 27 publications

(47 citation statements)

References 19 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Semantic segmentation architectures are typically trained on huge datasets with pixelwise annotations (e.g., the Cityscapes [5] or CamVid [1] datasets), which are highly expensive, time-consuming and error-prone to generate. To overcome this issue, semisupervised methods are emerging, trying to exploit weakly annotated data (e.g., with only image labels or only bounding boxes) [25,31,37,39,13,6,14,32] or completely unlabeled [24,29,15,31,19] data after a first stage of supervised training. In particular the works of [22,31] have paved the way respectively to adversarial learning approaches for the semantic segmentation task and to their application to semi-supervised learning.…”

Section: Related Workmentioning

confidence: 99%

“…In particular the works of [22,31] have paved the way respectively to adversarial learning approaches for the semantic segmentation task and to their application to semi-supervised learning. The recent approaches of [15,19] propose semi-supervised frameworks exploiting adversarial learning with a Fully Convolutional Discriminator (FCD) trying to distinguish the predicted probability maps from the ground truth segmentation distributions at pixel-level. These works targeted a scenario where the dataset is only partially labeled: in their settings, unlabeled data comes from the same dataset and shares the same domain data distribution of labeled data.…”

Section: Related Workmentioning

confidence: 99%

“…The first is a standard supervised cross-entropy loss exploiting ground truth annotations allowing to perform an initial supervised training phase on synthetic data. The second is an adversarial loss derived from previous methods [15,19] developed in the context of semi-supervised semantic segmentation (i.e., for dealing with datasets only partially annotated). In this framework, we exploited a fully convolutional discriminator which takes in input the semantic segmentation from the generator network and the ground truth segmentation maps and produces a pixel-level confidence map distinguishing between the two types of data.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Unsupervised Domain Adaptation for Semantic Segmentation of Urban Scenes

Biasetton

Michieli

Agresti

et al. 2019

2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW)

View full text Add to dashboard Cite

The semantic understanding of urban scenes is one of the key components for an autonomous driving system. Complex deep neural networks for this task require to be trained with a huge amount of labeled data, which is difficult and expensive to acquire. A recently proposed workaround is the usage of synthetic data, however the differences between real world and synthetic scenes limit the performance. We propose an unsupervised domain adaptation strategy to adapt a synthetic supervised training to real world data. The proposed learning strategy exploits three components: a standard supervised learning on synthetic data, an adversarial learning strategy able to exploit both labeled synthetic data and unlabeled real data and finally a self-teaching strategy working on unlabeled data only. The last component is guided by the segmentation confidence, estimated by the fully convolutional discriminator of the adversarial learning module, helping to further reduce the domain shift between synthetic and real data. Furthermore we weighted this loss on the basis of the class frequencies to enhance the performance on less common classes. Experimental results prove the effectiveness of the proposed strategy in adapting a segmentation network trained on synthetic datasets, like GTA5 and SYNTHIA, to a real dataset as Cityscapes.

show abstract

Section: Related Workmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Unsupervised Domain Adaptation for Semantic Segmentation of Urban Scenes

Biasetton

Michieli

Agresti

et al. 2019

2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW)

View full text Add to dashboard Cite

show abstract

“…The first component that controls the training is a standard cross-entropy loss exploiting ground truth annotations used to perform a supervised training on synthetic data. The second is an adversarial learning scheme similar to the ones used in works (e.g., [5], [6]) dealing with semi-supervised semantic segmentation (i.e., for dealing with partially annotated datasets). In particular, we exploited a fully convolutional discriminator which produces a pixel-level confidence map distinguishing between data produced by the generator (both from real or synthetic data) and the ground truth segmentation maps.…”

Section: Introductionmentioning

confidence: 99%

Adversarial Learning and Self-Teaching Techniques for Domain Adaptation in Semantic Segmentation

Michieli

Biasetton

Agresti

et al. 2020

IEEE Trans. Intell. Veh.

View full text Add to dashboard Cite

Deep learning techniques have been widely used in autonomous driving systems for the semantic understanding of urban scenes, however they need a huge amount of labeled data for training, which is difficult and expensive to acquire. A recently proposed workaround is to train deep networks using synthetic data, however the domain shift between real world and synthetic representations limits the performance. In this work a novel unsupervised domain adaptation strategy is introduced to solve this issue. The proposed learning strategy is driven by three components: a standard supervised learning loss on labeled synthetic data, an adversarial learning module that exploits both labeled synthetic data and unlabeled real data and finally a selfteaching strategy exploiting unlabeled data. The last component exploits a region growing framework guided by the segmentation confidence. Furthermore, we weighted this component on the basis of the class frequencies to enhance the performance on less common classes. Experimental results prove the effectiveness of the proposed strategy in adapting a segmentation network trained on synthetic datasets, like GTA5 and SYNTHIA, to real world datasets like Cityscapes and Mapillary.

show abstract

“…In this work, we develop a new data-driven deep spectral learning (DSL) method to enable highly robust and reliable sO2 estimation. By training a neural network to directly relate the spectral measurements to the corresponding independent sO2 labels, DSL bypasses the need for a rigid parametric model, similar to existing deep learning methods for solving optical inverse problems [33][34][35][36][37]. We show that DSL can be trained to be highly robust to multiple sources of variabilities in the experiments, including different devices, imaging protocols, speeds, and other possible longitudinal variations.…”

mentioning

confidence: 99%

Deep spectral learning for label-free optical imaging oximetry with uncertainty quantification

Liu¹,

Cheng

Tian

et al. 2019

Preprint

View full text Add to dashboard Cite

Measurement of blood oxygen saturation (sO2) by optical imaging oximetry provides invaluable insight into local tissue functions and metabolism. Despite different embodiments and modalities, all label-free optical imaging oximetry utilize the same principle of sO2-dependent spectral contrast from hemoglobin. Traditional approaches for quantifying sO2 often rely on analytical models that are fitted by the spectral measurements. These approaches in practice suffer from uncertainties due to biological variability, tissue geometry, light scattering, systemic spectral bias, and variations in experimental conditions. Here, we propose a new data-driven approach, termed deep spectral learning (DSL) for oximetry to be highly robust to experimental variations, and more importantly to provide uncertainty quantification for each sO2 prediction. To demonstrate the robustness and generalizability of DSL, we analyze data from two visible light optical coherence tomography (vis-OCT) setups across two separate in vivo experiments in rat retina. Predictions made by DSL are highly adaptive to experimental variabilities as well as the depth-dependent backscattering spectra. Two neural-network-based models are tested and compared with the traditional least-squares fitting (LSF) method. The DSL-predicted sO2 shows significantly lower mean-square errors than the LSF. For the first time, we have demonstrated en face maps of retinal oximetry along with pixel-wise confidence assessment. Our DSL overcomes several limitations in the traditional approaches and provides a more flexible, robust, and reliable deep learning approach for in vivo non-invasive label-free optical oximetry.

show abstract

Semi-Supervised Automatic Layer and Fluid Region Segmentation of Retinal Optical Coherence Tomography Images Using Adversarial Learning

Cited by 27 publications

References 19 publications

Unsupervised Domain Adaptation for Semantic Segmentation of Urban Scenes

Unsupervised Domain Adaptation for Semantic Segmentation of Urban Scenes

Adversarial Learning and Self-Teaching Techniques for Domain Adaptation in Semantic Segmentation

Deep spectral learning for label-free optical imaging oximetry with uncertainty quantification

Contact Info

Product

Resources

About