Navigation

A probabilistic approach to image resampling

Proposal for a Master Thesis

Topic:

A probabilistic approach to image resampling

Description:

Image processing often produces pixels that are not located on integer pixel positions. If an image is e.g. rotated, the pixels are located on non-integer positions after the rotation. We refer to these points as mesh. It is not desired that pixels are lying on a mesh as it cannot be stored efficiently. Hence, a method has to be found that interpolates the mesh pixels onto regular integer positions which we call grid. For interpolation various techniques already exist. Probabilistic approaches to interpolation are not very commonly used in image processing. Especially, Gaussian processes are not very important for interpolation in image processing at the moment.

Download

Professor:

Prof. Dr.-Ing. André Kaup

Supervisior:

Viktoria Heimann, Room 06.036, viktoria.heimann@fau.de

Prerequisites:

MATLAB, Digital Signal Processing, Basics in Image Processing, proficient in German or English

Available:

Immediately