Skip to main content

LUP Student Papers

LUND UNIVERSITY LIBRARIES

Nonlinear Interpolation from Video Images to High Quality Printers

Hansson, Anders (1989) In MSc Theses
Department of Automatic Control
Abstract
Recent high quality printers pose interesting questions on how to reproduce images taken by a video camera. A modern ink jet plotter has typically a resolution four times better than a good video camera. Pixel replication and bilinear interpolation give reasonable results, but have a tendency to blur the image. In contrast to these blurring phenomena, it is well known that the apparent visual quality is crucially dependent on how well the singularities, like edges, are reproduced. It is therefore intriguing to base the interpolation on a theoretical formulation that explicitly is based on image singularities. Classical potential theory fulfills these requirements, and the aim has been to start investigating the ability of this theory to... (More)
Recent high quality printers pose interesting questions on how to reproduce images taken by a video camera. A modern ink jet plotter has typically a resolution four times better than a good video camera. Pixel replication and bilinear interpolation give reasonable results, but have a tendency to blur the image. In contrast to these blurring phenomena, it is well known that the apparent visual quality is crucially dependent on how well the singularities, like edges, are reproduced. It is therefore intriguing to base the interpolation on a theoretical formulation that explicitly is based on image singularities. Classical potential theory fulfills these requirements, and the aim has been to start investigating the ability of this theory to capture intrinsic image properties. Computational algorithms have been searched for in two ways. One direction is to use numerical methods on the full problem. Another direction is to search for formulations approximating the full theory but having less computational complexity. Algorithms of both types are developed. They are nonlinear and make use of an edge-detection technique that is implemented as a least-square-minimization. An explicit description of images by means of a small number of primitives has been obtained. The nonlinear interpolation schemes have shown to preserve edges well. (Less)
Please use this url to cite or link to this publication:
author
Hansson, Anders
supervisor
organization
year
type
H3 - Professional qualifications (4 Years - )
subject
keywords
Image interpolation, Edge-detection, Potential theory, Ink jet plotter
publication/series
MSc Theses
report number
TFRT-5397
ISSN
0280-5316
language
English
id
8849363
date added to LUP
2016-03-26 11:46:39
date last changed
2016-03-26 11:46:39
@misc{8849363,
  abstract     = {{Recent high quality printers pose interesting questions on how to reproduce images taken by a video camera. A modern ink jet plotter has typically a resolution four times better than a good video camera. Pixel replication and bilinear interpolation give reasonable results, but have a tendency to blur the image. In contrast to these blurring phenomena, it is well known that the apparent visual quality is crucially dependent on how well the singularities, like edges, are reproduced. It is therefore intriguing to base the interpolation on a theoretical formulation that explicitly is based on image singularities. Classical potential theory fulfills these requirements, and the aim has been to start investigating the ability of this theory to capture intrinsic image properties. Computational algorithms have been searched for in two ways. One direction is to use numerical methods on the full problem. Another direction is to search for formulations approximating the full theory but having less computational complexity. Algorithms of both types are developed. They are nonlinear and make use of an edge-detection technique that is implemented as a least-square-minimization. An explicit description of images by means of a small number of primitives has been obtained. The nonlinear interpolation schemes have shown to preserve edges well.}},
  author       = {{Hansson, Anders}},
  issn         = {{0280-5316}},
  language     = {{eng}},
  note         = {{Student Paper}},
  series       = {{MSc Theses}},
  title        = {{Nonlinear Interpolation from Video Images to High Quality Printers}},
  year         = {{1989}},
}