Multi-penalty conditional random field approach to super-resolved reconstruction of optical coherence tomography images

Abstract

Improving the spatial resolution of Optical Coherence Tomography (OCT) images is important for the visualization and analysis of small morphological features in biological tissue such as blood vessels, membranes, cellular layers, etc. In this paper, we propose a novel reconstruction approach to obtaining super-resolved OCT tomograms from multiple lower resolution images. The proposed Multi-Penalty Conditional Random Field (MPCRF) method combines four different penalty factors (spatial proximity, first and second order intensity variations, as well as a spline-based smoothness of fit) into the prior model within a Maximum A Posteriori (MAP) estimation framework. Test carried out in retinal OCT images illustrate the effectiveness of the proposed MPCRF reconstruction approach in terms of spatial resolution enhancement, as compared to previously published super resolved image reconstruction methods. Visual assessment of the MPCRF results demonstrate the potential of this method in better preservation of fine details and structures of the imaged sample, as well as retaining the sharpness of biological tissue boundaries while reducing the effects of speckle noise inherent to OCT. Quantitative evaluation using imaging metrics such as Signal-to-Noise Ratio (SNR), Contrast to Noise Ratio (CNR), Equivalent Number of Looks (ENL), and Edge Preservation Parameter show significant visual quality improvement with the MPCRF approach. Therefore, the proposed MPCRF reconstruction approach is an effective tool for enhancing the spatial resolution of OCT images without the necessity for significant imaging hardware modifications.

Proliferative vitreoretinopathy: A new concept of disease pathogenesis and practical consequences

“…Even in recent papers these retinal changes have not been recognized (Rouberol and Chiquet, 2014). Nevertheless, there is hope that the use of new imaging technologies could add relevant information regarding the intraretinal changes (Boroomand et al, 2013).…”

Optical coherence tomography retinal image reconstruction via nonlocal weighted sparse representation

“…Also, to avoid motion artifacts from the fixation eye movements, the OCT clinical images are often captured at lower than nominal sampling rates. [2][3][4][5][6] Therefore, effective denoising and interpolation algorithms are necessary for automated or even manual OCT image analysis.…”

“…5,6,20 Classic interpolation algorithms produce unsatisfying results when the low-resolution image is noisy and they cannot recover features that are missed in the input noisy image itself. [4][5][6]20 Multiframe interpolation methods seems more plausible since they benefit from more information provided by multiple images. 2 However, they need motion estimation from images, which is prone to error even for noise-free images.…”

“…[23][24][25][26][27] Thus, several researchers have used sparse representation and closely related compressed sensing methods to reconstruct OCT data. [2][3][4][5][6][28][29][30] Recently, Fang et al 5,6 have tried to denoise and interpolate spectral domain OCT (SDOCT) images by proposing a fast coupled dictionary learning (DL) approach. Basically, this approach is based on the idea of shared sparse representations between LR and HR patches.…”

Enhancement of morphological and vascular features in OCT images using a modified Bayesian residual transform

“…Many computational approaches are also widely explored. Speckle noise reduction algorithms, such as Gaussian filter, linear least-square estimation [15], Bayesian estimation [16,17], conditional random field [18], adaptive median filtering [19], anisotropic diffusion [20] , estimate and compensate the image aberrations, therefore can enhance the visualization of features of interest. These approaches require good phase stability of the system and dense scanning pattern, so their transition to the clinic use systems is difficult and filed of view is limited in in-vivo imaging.…”