Projects
Image Denoising.
Description : In this project, we develop a Poisson denoising model based on the fractional-order total variation (FOTV) regularization.
Algorithms : To solve the problem eciently, we propose three numerical algorithms based on the Chambolle-Pock primal-dual method, a forward-backward splitting scheme, and the alternating direction method of multipliers (ADMM), each with guaranteed convergence.
Dataset: One synthetic image and five standard images, i.e., Cameraman, Penguin, Train, Mandrill, and Barbara, all of which are widely used in the image processing literature.
Image deblurring.
Description : In this project, we develop image deblurring model that deal with Poisson noise.
Algorithms : We develop an efficient algorithm based on the alternating direction method of multipliers (ADMM).
Dataset: we use the four standard testing images, named by “Cameraman,” “Galaxy,” “Phantom,” and “Shape”.
Blind Image deconvolution.
Description : In this project, we develop image deblurring model that deal with Poisson noise.
Algorithms : Expectation-maximization (EM) and ADMM.
Dataset: We conduct blind deconvolution using three testing images: Shape, Spine, and Satellite.
CT Image reconstruction.
Description : In this project, we develop image deblurring model that deal with Poisson noise.
Algorithms : ADMM.
Dataset: The Shepp–Logan phantom is a standard test image. It serves as the model of a human head in the development and testing of image reconstruction algorithms.
Image compression using SVD and NMF.
Description :
Algorithms :
Dataset:
PET Image Reconstruction.
Description :
Algorithms :
Dataset: