scholarly journals Diagnosis of central serous chorioretinopathy by deep learning analysis of en face images of choroidal vasculature: A pilot study

PLoS ONE ◽  
2021 ◽  
Vol 16 (6) ◽  
pp. e0244469
Author(s):  
Yukihiro Aoyama ◽  
Ichiro Maruko ◽  
Taizo Kawano ◽  
Tatsuro Yokoyama ◽  
Yuki Ogawa ◽  
...  
Keyword(s):  
Deep Learning ◽  
Active Region ◽  
Central Serous Chorioretinopathy ◽  
Small Data ◽  
Data Sets ◽  
Swept Source ◽  
The Neural Network ◽  
En Face ◽  
Choroidal Vasculature ◽  
The Mean

Purpose To diagnose central serous chorioretinopathy (CSC) by deep learning (DL) analyses of en face images of the choroidal vasculature obtained by optical coherence tomography (OCT) and to analyze the regions of interest for the DL from heatmaps. Methods One-hundred eyes were studied; 53 eyes with CSC and 47 normal eyes. Volume scans of 12×12 mm square were obtained at the same time as the OCT angiographic (OCTA) scans (Plex Elite 9000 Swept-Source OCT®, Zeiss). High-quality en face images of the choroidal vasculature of the segmentation slab of one-half of the subfoveal choroidal thickness were created for the analyses. The 100 en face images were divided into 80 for training and 20 for validation. Thus, we divided it into five groups of 20 eyes each, trained the remaining 80 eyes in each group, and then calculated the correct answer rate for each group by validation with 20 eyes. The Neural Network Console (NNC) developed by Sony and the Keras-Tensorflow backend developed by Google were used as the software for the classification with 16 layers of convolutional neural networks. The active region of the heatmap based on the feature quantity extracted by DL was also evaluated as the percentages with gradient-weighted class activation mapping implemented in Keras. Results The mean accuracy rate of the validation was 95% for NNC and 88% for Keras. This difference was not significant (P >0.1). The mean active region in the heatmap image was 12.5% in CSC eyes which was significantly lower than the 79.8% in normal eyes (P<0.01). Conclusions CSC can be automatically diagnosed by DL with high accuracy from en face images of the choroidal vasculature with different programs, convolutional layer structures, and small data sets. Heatmap analyses showed that the DL focused on the area occupied by the choroidal vessels and their uniformity. We conclude that DL can help in the diagnosis of CSC.

scholarly journals Diagnosis of central serous chorioretinopathy by deep learning analysis of en face images of choroidal vasculature

2020 ◽  
Author(s):  
Yukihiro Aoyama ◽  
Ichiro Maruko ◽  
Taizo Kawano ◽  
Tatsuro Yokoyama ◽  
Yuki Ogawa ◽  
...  
Keyword(s):  
Deep Learning ◽  
Active Region ◽  
Central Serous Chorioretinopathy ◽  
Small Data ◽  
Data Sets ◽  
Swept Source ◽  
Face Images ◽  
The Neural Network ◽  
En Face ◽  
Choroidal Vasculature

Purpose To classify central serous chorioretinopathy (CSC) by deep learning (DL) analyses of en face images of choroidal vasculature obtained by optical coherence tomography (OCT) and to analyze the regions of interest for DL from heatmaps. Methods One-hundred eyes were studied; 53 eyes with CSC and 47 normal eyes. Volume scans of 12×12 mm square were obtained at the same time as the OCT angiographic (OCTA) scans (Plex Elite 9000 Swept-Source OCT®, Zeiss). High-quality en face images of the choroidal vasculature of the segmentation slab of one-half of the subfoveal choroidal thickness were created for the analyses. The entire 100 en face images were divided into 80 for training (100 times) and 20 for validation. The Neural Network Console (NNC) developed by Sony and the Keras-Tensorflow backend and developed by Google were used as the software for the classification with 16 layers of convolutional neural networks. The active region of the heatmap based on the feature quantity extracted by DL was also evaluated as the percentages with gradient-weighted class activation mapping implementation in Keras. Results In the 20 eyes used for validation including 8 eyes with CSC, the accuracy rate of the validation was 100% (20/20) for NNC and 95% (19/20) for Keras. This difference was not significant ( P =0.33). The mean active region in the heatmap image was 12.5% in CSC eyes which was significantly lower than the 79.8% in normal eyes ( P <0.01). Conclusions CSC can be automatically classified with high accuracy from en face images of the choroidal vasculature by DLs with different programs, convolutional layer structures, and small data sets. Heatmap analyses showed that DL focused on the area occupied by the choroidal vessels and their uniformity. We conclude that DL can help in the diagnosis of CSC.

En-face choroidal vascularity in central serous chorioretinopathy

2020 ◽  
pp. 112067212090871 ◽  
Author(s):  
Raymond Lai-Man Wong ◽  
Sumit Randhir Singh ◽  
Mohammed Abdul Rasheed ◽  
Abhilash Goud ◽  
Gunjan Chhablani ◽  
...  
Keyword(s):  
Optical Coherence Tomography ◽  
Central Serous Chorioretinopathy ◽  
Optical Coherence ◽  
Chronic Central Serous Chorioretinopathy ◽  
Swept Source ◽  
Chronic Group ◽  
Acute Group ◽  
Luminal Area ◽  
En Face ◽  
The Mean

Purpose: To evaluate the choroidal vascularity index of eyes for acute and chronic central serous chorioretinopathy patients using swept-source optical coherence tomography generated en-face scans. Methods: This was a retrospective study, in which slabs of en-face optical coherence tomography scans, at 5 μm intervals, spanning from the retina to choroid, were binarized using a validated algorithm to calculate choroidal vascularity index. The choroidal vascularity index was defined as the ratio between the choroidal vascular luminal area and the total choroidal area. Choroidal vascularity index was calculated for all the slabs of every subject in both the groups. Results: A total of 30 eyes for each acute and chronic central serous chorioretinopathy groups were recruited. The mean choroidal vascularity index of the acute group was 45.21% ± 2.25% at the choriocapillaris, which increased to the maximal value of 48.35% ± 2.06% at 75% depth of the choroidal thickness and 45.31% ± 3.27% at the choroidoscleral interface; whereas for the chronic group, the mean choroidal vascularity index was 44.76% ± 2.60% at the choriocapillaris, which maximized at 50% choroidal depth (48.70% ± 1.32%) and then returned to 45.41% ± 6.02% at the choroidoscleral interface. Conclusion: For both groups, the choroidal vascularity index increased from choriocapillaris to maximum values at mid-choroid and returned to almost the choriocapillaris value at the choroidoscleral interface.

scholarly journals Imaging the Deep Choroidal Vasculature Using Spectral Domain and Swept Source Optical Coherence Tomography Angiography

2018 ◽  
Vol 2 (3) ◽  
pp. 146-154 ◽  
Author(s):  
J. Daniel Diaz ◽  
Jay C. Wang ◽  
Patrick Oellers ◽  
Inês Lains ◽  
Lucia Sobrin ◽  
...  
Keyword(s):  
Blood Flow ◽  
Optical Coherence Tomography Angiography ◽  
Choroidal Blood Flow ◽  
Swept Source ◽  
Blood Flow Signal ◽  
En Face ◽  
Rpe Atrophy ◽  
Choroidal Vasculature ◽  
Choroidal Vessels ◽  
En Face Imaging

Purpose: To evaluate the deeper choroidal vasculature in eyes with various ocular disorders using spectral domain (SD) optical coherence tomography angiography (OCTA) and swept source (SS) OCTA. Methods: Patients underwent OCTA imaging with either SD-OCTA (Zeiss Cirrus Angioplex or Optovue AngioVue) or SS-OCTA (Topcon Triton). Retinal pigment epithelium (RPE) integrity, structural visualization of deep choroidal vessels on en face imaging, and OCTA of deep choroidal blood flow signal were analyzed. Choroidal blood flow was deemed present if deeper choroidal vessels appeared bright after appropriate segmentation. Results: Structural visualization of choroidal vessels was feasible in all eyes by en face imaging. In both SD-OCTA and SS-OCTA, choroidal blood flow signal was present in all eyes with overlying RPE atrophy (100% of eyes with RPE atrophy, 28.6% of all imaged eyes, P < .001). Conclusions: While choroidal vessels can be visualized anatomically in all eyes by en face imaging, choroidal blood flow detection in deep choroidal vessel is largely restricted to areas with overlying RPE atrophy. Intact RPE acts as a barrier for reliable detection of choroidal flow using current OCTA technology, inhibiting evaluation of flow in deeper choroidal vessels in most eyes.

Pattern-Based Deep Learning Method to Extract Information from the Log Dataset

2021 ◽  
pp. 2150296
Author(s):  
Xi Li ◽  
Ting Wang ◽  
Shexiong Wang
Keyword(s):  
Neural Network ◽  
Deep Learning ◽  
Data Sets ◽  
Learning Method ◽  
Statistical Features ◽  
Log Data ◽  
The Neural Network ◽  
Testing Data ◽  
Extract Information

It draws researchers’ attentions how to make use of the log data effectively without paying much for storing them. In this paper, we propose pattern-based deep learning method to extract the features from log datasets and to facilitate its further use at the reasonable expense of the storage performances. By taking the advantages of the neural network and thoughts to combine statistical features with experts’ knowledge, there are satisfactory results in the experiments on some specified datasets and on the routine systems that our group maintains. Processed on testing data sets, the model is 5%, at least, more likely to outperform its competitors in accuracy perspective. More importantly, its schema unveils a new way to mingle experts’ experiences with statistical log parser.

scholarly journals Quantitative assessment of choriocapillaris flow deficits in diabetic retinopathy: A swept-source optical coherence tomography angiography study

PLoS ONE ◽  
2020 ◽  
Vol 15 (12) ◽  
pp. e0243830
Author(s):  
Yining Dai ◽  
Hao Zhou ◽  
Qinqin Zhang ◽  
Zhongdi Chu ◽  
Lisa C. Olmos de Koo ◽  
...  
Keyword(s):  
Optical Coherence Tomography ◽  
Diabetic Retinopathy ◽  
Healthy Subjects ◽  
Optical Coherence Tomography Angiography ◽  
Optical Coherence ◽  
Swept Source ◽  
Significant Difference ◽  
En Face ◽  
The Mean

Purpose To quantitatively assess choriocapillaris (CC) flow deficits in eyes with diabetic retinopathy (DR) using swept-source optical coherence tomography angiography (SS-OCTA). Methods Diabetic subjects with different stages of DR and age-matched healthy subjects were recruited and imaged with SS-OCTA. The en face CC blood flow images were generated using previously published and validated algorithms. The percentage of CC flow deficits (FD%) and the mean CC flow deficit size were calculated in a 5-mm-diameter circle centered on the fovea from the 6×6-mm scans. Results Forty-five diabetic subjects and 27 control subjects were included in the study. The CC FD% in diabetic eyes was on average 1.4-fold greater than in control eyes (12.34±4.14% vs 8.82±2.61%, P < 0.001). The mean CC FD size in diabetic eyes was on average 1.4-fold larger than in control eyes (2151.3± 650.8μm2 vs 1574.4±255.0 μm2, P < 0.001). No significant difference in CC FD% or mean CC FD size was observed between eyes with nonproliferative DR and eyes with proliferative DR (P = 1.000 and P = 1.000, respectively). Conclusions CC perfusion in DR can be objectively and quantitatively assessed with FD% and FD size. In the macular region, both CC FD% and CC FD size are increased in eyes with DR. SS-OCTA provides new insights for the investigations of CC perfusion status in diabetes in vivo.

scholarly journals Deep learning for Chinese NOx emission inversion and the integration of in situ observations: a case study on the COVID-19 pandemic

2021 ◽  
Author(s):  
Tai-Long He ◽  
Dylan Jones ◽  
Kazuyuki Miyazaki ◽  
Kevin Bowman ◽  
Zhe Jiang ◽  
...  
Keyword(s):  
Deep Learning ◽  
Reanalysis Data ◽  
Data Sets ◽  
Computationally Efficient ◽  
Transport Models ◽  
The Neural Network ◽  
In Situ Observations ◽  
Typical Variation

&lt;p&gt;The COVID-19 pandemic led to the lockdown of over one-third of Chinese cities in early 2020. Observations have shown significant reductions of atmospheric abundances of NO&lt;sub&gt;2&lt;/sub&gt; over China during this period. This change in atmospheric NO&lt;sub&gt;2&lt;/sub&gt; implies a dramatic change in emission of NO&lt;sub&gt;x&lt;/sub&gt;, which provides a unique opportunity to study the response of the chemistry of the atmospheric to large reductions in anthropogenic emissions. We use a deep learning (DL) model to quantify the change in surface emissions of NO&lt;sub&gt;x&lt;/sub&gt; in China that are associated with the observed changes in atmospheric NO&lt;sub&gt;2&lt;/sub&gt; during the lockdown period. Compared to conventional data assimilation systems, deep neural networks are free of the potential errors associated with parameterized subgrid-scale processes. Furthermore, they are not susceptible to the chemical errors typically found in atmospheric chemical transport models. The neural-network-based approach also offers a more computationally efficient means of inverse modeling of NO&lt;sub&gt;x&lt;/sub&gt; emissions at high spatial resolutions. Our DL model is trained using meteorological predictors and reanalysis data of surface NO&lt;sub&gt;2&lt;/sub&gt; from 2005 to 2017. The evaluation is conducted using in-situ measurements of NO&lt;sub&gt;2&lt;/sub&gt; in 2019 and 2020. The Baidu 'Qianxi' migration data sets are used to evaluate the model's performance in capturing the typical variation in Chinese NOx emissions during the Chinese New Year holidays. The TROPOMI-derived TCR-2 chemical reanalysis is used to evaluate the DL analysis in 2020. We show that the DL-based approach is able to better reproduce the variation in anthropogenic NO&lt;sub&gt;x&lt;/sub&gt; emissions and capture the reduction in Chinese NO&lt;sub&gt;x&lt;/sub&gt; emissions during the period of the COVID-19 pandemic.&lt;/p&gt;

scholarly journals Visualization of Choriocapillaris and Choroidal Vasculature in Healthy Eyes With En Face Swept-Source Optical Coherence Tomography Versus Angiography

2018 ◽  
Vol 7 (6) ◽  
pp. 25 ◽  
Author(s):  
Jay C. Wang ◽  
Inês Laíns ◽  
Rebecca F. Silverman ◽  
Lucia Sobrin ◽  
Demetrios G. Vavvas ◽  
...  
Keyword(s):  
Optical Coherence Tomography ◽  
Optical Coherence ◽  
Swept Source ◽  
En Face ◽  
Choroidal Vasculature ◽  
Healthy Eyes

scholarly journals Deep transfer learning for aortic root dilation identification in 3D ultrasound images

2018 ◽  
Vol 4 (1) ◽  
pp. 71-74 ◽  
Author(s):  
Jannis Hagenah ◽  
Mattias Heinrich ◽  
Floris Ernst
Keyword(s):  
Deep Learning ◽  
Transfer Learning ◽  
Aortic Root ◽  
Classification Accuracy ◽  
3D Ultrasound ◽  
Small Data ◽  
Data Sets ◽  
Ultrasound Images ◽  
Small Data Sets ◽  
Learned Features

AbstractPre-operative planning of valve-sparing aortic root reconstruction relies on the automatic discrimination of healthy and pathologically dilated aortic roots. The basis of this classification are features extracted from 3D ultrasound images. In previously published approaches, handcrafted features showed a limited classification accuracy. However, feature learning is insufficient due to the small data sets available for this specific problem. In this work, we propose transfer learning to use deep learning on these small data sets. For this purpose, we used the convolutional layers of the pretrained deep neural network VGG16 as a feature extractor. To simplify the problem, we only took two prominent horizontal slices throgh the aortic root, the coaptation plane and the commissure plane, into account by stitching the features of both images together and training a Random Forest classifier on the resulting feature vectors. We evaluated this method on a data set of 48 images (24 healthy, 24 dilated) using 10-fold cross validation. Using the deep learned features we could reach a classification accuracy of 84 %, which clearly outperformed the handcrafted features (71 % accuracy). Even though the VGG16 network was trained on RGB photos and for different classification tasks, the learned features are still relevant for ultrasound image analysis of aortic root pathology identification. Hence, transfer learning makes deep learning possible even on very small ultrasound data sets.

Sign in / Sign up

Export Citation Format

Share Document