Image deconvolution via efficient sparsifying transform learning Hizli Seyreklȩstirici Dönüsüm Öǧrenme ile Görüntü Ters Evrisimi

Date

2018-07-05

Author

Akyon, Fatih
Kamaci, Ulas
Öktem, Sevinç Figen

Metadata

Show full item record

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

Item Usage Stats

42
views

0
downloads

Image deconvolution is one of the most frequently encountered inverse problems in imaging. Since natural images can be modeled sparsely in some transform domain, sparsity priors have been shown to effectively regularize these problems and enable high-quality reconstructions. In this paper, we develop a data-adaptive sparse image reconstruction approach for image deconvolution based on transform learning. Our framework adaptively learns a patch-based sparsifying transform and simultaneously reconstructs the image from its noisy blurred measurement. This is achieved by solving the resulting optimization problem using an alternating minimization algorithm which has closed-form and efficient update steps. The performance of the developed algorithm is illustrated for an application in optical imaging by considering different optical blurs and noise levels. The results demonstrate that the developed method not only improves the reconstruction quality compared to the total-variation based approach, but also is fast.

Subject Keywords

Image deconvolution, Transform learning, Sparsity-based reconstruction, Optical imaging, Alternating minimization

URI

https://hdl.handle.net/11511/43357

DOI

https://doi.org/10.1109/siu.2018.8404295

Collections

Department of Electrical and Electronics Engineering, Conference / Seminar

Suggestions

OpenMETU
Core

Efficient algorithms for convolutional inverse problems in multidimensional imaging Doğan, Didem; Öktem, Figen S.; Department of Electrical and Electronics Engineering (2020) Computational imaging is the process of indirectly forming images from measurements using image reconstruction algorithms that solve inverse problems. In many inverse problems in multidimensional imaging such as spectral and depth imaging, the measurements are in the form of superimposed convolutions related to the unknown image. In this thesis, we first provide a general formulation for these problems named as convolutional inverse problems, and then develop fast and efficient image reconstruction algorith...
DEEP LEARNING-BASED UNROLLED RECONSTRUCTION METHODS FOR COMPUTATIONAL IMAGING Bezek, Can Deniz; Öktem, Sevinç Figen; Department of Electrical and Electronics Engineering (2021-9-08) Computational imaging is the process of forming images from indirect measurements using computation. In this thesis, we develop deep learning-based unrolled reconstruction methods for various computational imaging modalities. Firstly, we develop two deep learning-based reconstruction methods for diffractive multi-spectral imaging. The first approach is based on plug-and-play regularization with deep denoisers whereas the second one is an end-to-end learned reconstruction based on unrolling. Secondly, we con...
Cognitive aspects of image zooming Öncül, Gaye; Tarı, Sibel; Tekman, Hasan Gürkan; Department of Cognitive Sciences (2002) In this work, digital image zooming methods and image distortion metrics are examined in terms of a quality criterion based on image appearance. Neurophysiological and psychophysical foundations of perception driven image distortion metrics are presented. Results of a comparative study based on both perceptual distortion metrics and conventional distortion metrics for standard zooming methods are discussed. Open problems in image quality definition as well as image enlargement methods are presented and impr...
Fusion based resolution enhancement in hyperspectral images Hiperspektral Görüntulerde Kaynaştirma Temelli Çözünürlük Artirimi IRMAK, Hasan; Akar, Gözde; Yuksel, Seniha Esen (2017-05-18) Increasing the low spatial resolution of hyperspectcal images (HSIs) improves the performance of applications in which the HSIs are used. In this study, a fusion based method is proposed to increase the resolution of IISIs. In the proposed method, low resolution (LR) HSI is fused with the high resolution (HR) RGB image to obtain the HR HSI. In this approach, instead of using the spectral images as in the conventional methods, RGB image is used with the abundance maps of the HSI estimated from the linear unm...
An Investigation on hyperspectral image classifiers for remote sensing Özdemir, Okan Bilge; Çetin, Yasemin; Department of Information Systems (2013) Hyperspectral image processing is improved by the capabilities of multispectral image processing with high spectral resolution. In this thesis, we explored hyperspectral classification with Support Vector Machines (SVM), Maximum Likelihood (ML) and KNearest Neighborhood algorithms. We analyzed the effect of training data on classification accuracy. For this purpose, we implemented three different training data selection methods; first N sample selection, randomly N sample selection and uniformly N sample se...

Citation Formats

F. Akyon, U. Kamaci, and S. F. Öktem, “Image deconvolution via efficient sparsifying transform learning Hizli Seyreklȩstirici Dönüsüm Öǧrenme ile Görüntü Ters Evrisimi,” 2018, Accessed: 00, 2020. [Online]. Available: https://hdl.handle.net/11511/43357.