Super-Resolution OCT Based on α-Stable Distributions and Sparse Representations

Research output: Chapter in Book/Report/Conference proceedingConference Contribution (Conference Proceeding)

61 Downloads (Pure)


We present a new approach to single-image super-resolution in Optical Coherence Tomography (OCT) images. Indeed, although OCT is the one in-vivo retinal imaging modality offering the highest resolution, this is still far below what microscopy techniques can achieve, albeit ex-vivo. In this work we investigate a non-convex regularization technique using a multivariate generalization of the minimax-concave (GMC) scheme and a forward-backward splitting (FBS) algorithm. Based on the observation that sparse representations of OCT images are heavytailed, an -stable dictionary is employed. The resulting algorithm is tested on real OCT retinal images of murine eyes. Significant deblurring and general quality enhancement is noticed and in most cases our method provides the best results both objectively and subjectively.
Original languageEnglish
Title of host publicationProceedings of the International BASP Frontiers Workshop 2019
PublisherHeriot-Watt University
Number of pages1
Publication statusPublished - 3 Feb 2019
EventInternational Biomedical and Astronomical Signal Processing Frontiers workshop 2019 - Villars-sur-Ollon, Switzerland
Duration: 3 Feb 20198 Feb 2019
Conference number: 5

Publication series

NameInternational BASP Frontiers workshop


WorkshopInternational Biomedical and Astronomical Signal Processing Frontiers workshop 2019
Abbreviated titleBASP Frontiers 2019
Internet address


Dive into the research topics of 'Super-Resolution OCT Based on α-Stable Distributions and Sparse Representations'. Together they form a unique fingerprint.

Cite this