Browsing by Author "Kamasak, Mustafa E."

Now showing 1 - 5 of 5

Unknown
A discretized tomographic image reconstruction based upon total variation regularization
(Elsevier Sci Ltd, The Boulevard, Langford Lane, Kidlington, Oxford Ox5 1Gb, Oxon, England, 2017-09) Demircan Türeyen, Ezgi; Kamasak, Mustafa E.; 237397
Tomographic image reconstruction problem has an ill-posed nature like many other linear inverse problems in the image processing domain. Discrete tomography (DT) techniques are developed to cope with this drawback by utilizing the discreteness of an image. Discrete algebraic reconstruction technique (DART) is a DT technique that alternates between an inversion stage, employed by the algebraic reconstruction methods (ARM), and a discretization (i.e. segmentation) stage. Total variation (TV) minimization is another popular technique that deals with the ill-posedness by exploiting the piece-wise constancy of the image and basically requires to solve a convex optimization problem. In this paper, we propose an algorithm which also performs the successive sequences of inversion and discretization, but it estimates the continuous reconstructions under TV-based regularization instead of using ARM. Our algorithm incorporates the DART's idea of reducing the number of unknowns through the subsequent iterations, with a 1-D TV-based setting. As a second contribution, we also suggest a procedure to be able to select the segmentation parameters automatically which can be applied when the gray levels (corresponding to the different densities in the scanned object) are not known a priori. We performed various experiments using different phantoms, to show the proposed algorithm reveals better approximations when compared to DART, as well as three other continuous reconstruction techniques. While investigating the performances, we considered limited number of projections, limited-view, noisy projections and lack of prior knowledge on gray levels scenarios. (C) 2017 Elsevier Ltd. All rights reserved.
Unknown
Adaptive Direction-Guided Structure Tensor Total Variation
(Elsevier, 2021) Kamasak, Mustafa E.; TÜREYEN, EZGİ DEMİRCAN
Direction-guided structure tensor total variation (DSTV) is a recently proposed regularization term that aims at increasing the sensitivity of the structure tensor total variation (STV) to the changes towards a predetermined direction. Despite of the plausible results obtained on the uni-directional images, the DSTV model is not applicable to the arbitrary (multi-directional and/or partly nondirectional) images. In this study, we build a two-stage denoising framework that brings adaptivity to the DSTV based denoising. We design a DSTV-like alternative to STV, which encodes the first-order information within a local neighborhood under the guidance of spatially varying directional descriptors (i.e., orientation and the dose of anisotropy). In order to estimate those descriptors, we propose an efficient preprocessor that captures the local geometry based on the structure tensor. Through the extensive experiments, we demonstrate how beneficial the involvement of the directional information in STV is, by comparing the proposed method with the state-of-the-art analysis-based denoising models, both in terms of quality and computational efficiency.
Unknown
Image Reconstruction from Sparse Samples Using Directional Total Variation Minimization
(IEEE, 345 E 47Th St, New York, Ny 10017 USA, 2016) Demircan Türeyen, Ezgi; Kamasak, Mustafa E.; Bayram, İlker; 237397; 190955
This paper considers reconstruction of missing pixels and formulates the problem under directional total variation (DTV) regularization. In order to devise an algorithm, forward-backward splitting method is used as a convex optimization tool, in conjunction with a fast projected gradient-based algorithm. The results are compared with the results of TV-based setting, and the utility of using DTV is shown in terms of accuracy, when an image with a dominant direction is the case.
Unknown
Nonlocal Adaptive Direction-Guided Structure Tensor Total Variation for Image Recovery
(Springer London Ltd., 2021) TÜREYEN, EZGİ DEMİRCAN; Kamasak, Mustafa E.
A common strategy in variational image recovery is utilizing the nonlocal self-similarity property, when designing energy functionals. One such contribution is nonlocal structure tensor total variation (NLSTV), which lies at the core of this study. This paper is concerned with boosting the NLSTV regularization term through the use of directional priors. More specifically, NLSTV is leveraged so that, at each image point, it gains more sensitivity in the direction that is presumed to have the minimum local variation. The actual difficulty here is capturing this directional information from the corrupted image. In this regard, we propose a method that employs anisotropic Gaussian kernels to estimate directional features to be later used by our proposed model. The experiments validate that our entire two-stage framework achieves better results than the NLSTV model and two other competing local models, in terms of visual and quantitative evaluation.
Unknown
Restoring Fluorescence Microscopy Images by Transfer Learning From Tailored Data
(IEEE-Inst Electrical Electronics Engineers Inc., 2022) AKBULUT, FATMA PATLAR; TÜREYEN, EZGİ DEMİRCAN; Kamasak, Mustafa E.
In fluorescence microscopy imaging, noise is a very usual phenomenon. To some extent, it can be suppressed by increasing the amount of the photon exposure; however, it is not preferable since this may not be tolerated by the subjected specimen. Thus, a sophisticated computational method is needed to denoise each acquired micrograph, so that they become more adequate for further feature extraction and image analysis. However, apart from the difficulties of the denoising problem itself, one main challenge is that the absence of the ground-truth images makes the data-driven techniques less applicable. In order to tackle this challenge, we suggest to tailor a dataset by handpicking images from unrelated source datasets. Our tailoring strategy involves exploring some low-level view-based features of the candidate images, and their similarities to those of the fluorescence microscopy images. We pretrain and fine-tune the well-known feed-forward denoising convolutional neural networks (DnCNNs) on our tailored dataset and a very limited amount of fluorescence images, respectively to ensure both the diversity and the content-awareness. The quantitative and visual experimentation show that our approach is able to curate a dataset, which is significantly superior to the arbitrarily chosen source images, and well-approximates to the fluorescence images. Moreover, the combination of the tailored dataset with a few fluorescence data through the use of fine-tuning offers a good balance between the generalization capability and the content-awareness, on the majority of considered scenarios.