RT Dissertation/Thesis
T1 Visibility Recovery on Images Acquired in Attenuating Media. Application to Underwater, Fog, and Mammographic Imaging
A1 Galdran, Adrian
AB When acquired in attenuating media, digital images often suffer from aparticularly complex degradation that reduces their visual quality, hinderingtheir suitability for further computational applications, or simplydecreasing the visual pleasantness for the user. In these cases, mathematicalimage processing reveals itself as an ideal tool to recover someof the information lost during the degradation process. In this dissertation,we deal with three of such practical scenarios in which this problematicis specially relevant, namely, underwater image enhancement, fogremoval and mammographic image processing. In the case of digital mammograms,X-ray beams traverse human tissue, and electronic detectorscapture them as they reach the other side. However, the superpositionon a bidimensional image of three-dimensional structures produces lowcontrastedimages in which structures of interest suffer from a diminishedvisibility, obstructing diagnosis tasks. Regarding fog removal, the lossof contrast is produced by the atmospheric conditions, and white colourtakes over the scene uniformly as distance increases, also reducing visibility.For underwater images, there is an added difficulty, since colour is notlost uniformly; instead, red colours decay the fastest, and green and bluecolours typically dominate the acquired images. To address all these challenges,in this dissertation we develop new methodologies that rely on: a)physical models of the observed degradation, and b) the calculus of variations.Equipped with this powerful machinery, we design novel theoreticaland computational tools, including image-dependent functional energiesthat capture the particularities of each degradation model. These energiesare composed of different integral terms that are simultaneouslyminimized by means of efficient numerical schemes, producing a clean,visually-pleasant and useful output image, with better contrast and increasedvisibility. In every considered application, we provide comprehensivequalitative (visual) and quantitative experimental results to validateour methods, confirming that the developed techniques outperform otherexisting approaches in the literature.
PB UPV-EHU
YR 2015
FD 2015-12-17
LK http://hdl.handle.net/11556/199
UL http://hdl.handle.net/11556/199
LA eng
DS TECNALIA Publications
RD 27 jul 2024