Improve synthetic retinal OCT images with present of pathologies and textural information

Varnousfaderani, Ehsan Shahrian; Vogl, Wolf‐Dieter; Wu, Jing; Gerendas, Bianca S.; Simader, Christian; Waldstein, Sebastian M.; Schmidt‐Erfurth, Ursula

doi:10.1117/12.2217399

Cited by 4 publications

(11 citation statements)

References 6 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Unlike other studies on OCT synthesis [16, 17], which only rely on segmented data from internal limiting membrane (ILM), outer segment layer (OSL) and retinal pigment epithelium (RPE), the proposed method is more detailed by incorporating information from nine layers of retina. ASM method is also novel in producing new OCT shape parameters.…”

Section: Resultsmentioning

confidence: 99%

Automatic Production of Synthetic Labeled OCT Images Using Active Shape Model

Danesh

Maghooli

Kafieh

et al. 2020

Preprint

View full text Add to dashboard Cite

The challenge of limited labeled data in the field of medical imaging and the need for large number of labeled data for training machine learning algorithms, and to measure the performance of image processing algorithms increases the demand to use synthetic images. The purpose of this paper is to construct synthetic and labeled Optical Coherence Tomography (OCT) data to solve the problems like having access to the accurate labeled data and evaluating the processing algorithms. In this study, a modified active shape model is used which considers the anatomical features of available images such as number and thickness of the layers and their associated brightness, the retinal blood vessels, and shadow information with wise consideration of speckle noise. The algorithm is also able to provide different datasets with varying noise level. The validity of our method for synthesis of retinal images is measured by two methods (qualitative assessment and quantitative analysis).

show abstract

Section: Resultsmentioning

confidence: 99%

Automatic Production of Synthetic Labeled OCT Images Using Active Shape Model

Danesh

Maghooli

Kafieh

et al. 2020

Preprint

View full text Add to dashboard Cite

show abstract

“…The B-scans of five original noisy OCT volumes are shown in Fig7.a. The generated synthetic noisefree and noisy OCT volume using our proposed method in 15 The standard errors of the Mean Structural Similarity Index (MSSIM) 8 and Effective Number of Looks (ENL) 8 are used to evaluate the performance of the denoising methods on the generated set of synthetic OCT volumes (Fig7). The ENL is defined as…”

Section: Quantitative Evaluation Of Denoising Methodsmentioning

confidence: 99%

“…However, this method is limited to non-pathological OCT data, and pathologies such as cysts are simply smoothed away. We proposed in 15 an enhancement of their method by applying smoothness functions locally. In addition, pathological cyst regions are added from manual cyst annotations of the real OCT scan, on which the synthetic scan is based on.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Geodesic denoising for optical coherence tomography images

Varnousfaderani

Vogl²,

Wu³

et al. 2016

SPIE Proceedings

Self Cite

View full text Add to dashboard Cite

Optical coherence tomography (OCT) is an optical signal acquisition method capturing micrometer resolution, crosssectional three-dimensional images. OCT images are used widely in ophthalmology to diagnose and monitor retinal diseases such as age-related macular degeneration (AMD) and Glaucoma. While OCT allows the visualization of retinal structures such as vessels and retinal layers, image quality and contrast is reduced by speckle noise, obfuscating small, low intensity structures and structural boundaries. Existing denoising methods for OCT images may remove clinically significant image features such as texture and boundaries of anomalies.In this paper, we propose a novel patch based denoising method, Geodesic Denoising. The method reduces noise in OCT images while preserving clinically significant, although small, pathological structures, such as fluid-filled cysts in diseased retinas. Our method selects optimal image patch distribution representations based on geodesic patch similarity to noisy samples. Patch distributions are then randomly sampled to build a set of best matching candidates for every noisy sample, and the denoised value is computed based on a geodesic weighted average of the best candidate samples.Our method is evaluated qualitatively on real pathological OCT scans and quantitatively on a proposed set of ground truth, noise free synthetic OCT scans with artificially added noise and pathologies. Experimental results show that performance of our method is comparable with state of the art denoising methods while outperforming them in preserving the critical clinically relevant structures.

show abstract

“…The primary methods for producing medical images were digital phantoms that followed simple mathematical models of human anatomy [16]. Several studies exist for producing synthetic retinal fundus images [15,17,18], CT scans [19,20], and OCT images [21][22][23][24] and several studies used synthetic data to compare and measure the performance of image processing algorithms [25][26][27]. The available literature on the production of synthetic OCT images is categorised in Table 1.…”

Section: Introductionmentioning

confidence: 99%

Automatic production of synthetic labelled OCT images using an active shape model

Danesh

Maghooli

Dehghani

et al. 2020

IET Image Processing

View full text Add to dashboard Cite

Limited labelled data is a challenge in the field of medical imaging and the need for a large number of them is paramount for the training of machine learning algorithms, as well as measuring the performance of image processing algorithms. The purpose of this study is to construct synthetic and labelled optical coherence tomography (OCT) data to solve the problems of having access to accurately labelled data and evaluating the processing algorithms. In this study, a modified active shape model is used which considers the anatomical features of available images such as the number and thickness of the layers as well as their associated brightness, the location of retinal blood vessels and shadow information with respect to speckle noise. The algorithm is also able to provide different data sets with the varying noise level. The validity of the proposed method for the synthesis of retinal images is measured by two methods (qualitative assessment and quantitative analysis).

show abstract

Improve synthetic retinal OCT images with present of pathologies and textural information

Cited by 4 publications

References 6 publications

Automatic Production of Synthetic Labeled OCT Images Using Active Shape Model

Automatic Production of Synthetic Labeled OCT Images Using Active Shape Model

Geodesic denoising for optical coherence tomography images

Automatic production of synthetic labelled OCT images using an active shape model

Contact Info

Product

Resources

About