ידידה חיימס (הנדסת חשמל, טכניון)
יום שלישי, 8.8.2017, 11:30
חדר 815, בניין מאייר, הפקולטה להנדסת חשמל
The Total Variation (TV) framework has been shown to give a good scale-space representation for many purposes. Much has been done in showing the uses of TV for denoising, deconvolution and other spectral analysis tasks both for one-dimensional signals and for images, mainly for amplitude and grayscale images. In the field of colour imaging and compression, it is known that chromatic channels should be sampled and transmitted at a lower rate than that of the intensity. This is used in the vast most of cameras imaging systems. Such images then undergo a process of interpolation.
Most existing methods perform a form of low pass filtering and hence cause some loss of spatia
l frequencies in the image. In this work, we develop and suggest a method for performing such interpolation with edge preserving properties and hence conservation of high spatial frequencies in the image using the Total Variation flow. Our method interpolates based only on a single best neighbor and achieves comparable results to those of well-known methods. The results prove the ability to estimate the connection between unknown pixels and known pixels while assuming very little on the pattern of the sampling both in space and across the channels. Such framework could be later used as an immediate application for super-resoltion but also for inpainting and compression problems.
M.Sc. research under the supervision of Prof. Moshe Porat.