|
|
Picture correction
Sedlo, Petr ; Horák, Karel (referee) ; Richter, Miloslav (advisor)
This work is focussed on different unwanted effects that damage a digitally acquired image. An image acquired using a digital camera can suffer from a series of defects. First, the noise is concerned. It occures due to discretisation. Next, there are the defects including vignetting, radial image distortion, chromatic aberration. In my thesis I also worked on backlight image compensation and panorama creation from an image sequence. For backlight images we need to increase the brightness in the dark parts of the image, so that the details became more visible. At panorama creation we are careful about brightness evenness in partial images, which is usually not conserved and consequently we try to find a suitable boundary-line to connect images and smooth it. The fundamental part of my work was writing the algorithms that are able to compensate these defects at least partly. For this aim I took use of the Matlab environment.
|
|
|
Simulation of Lens Features
Kučiš, Michal ; Seeman, Michal (referee) ; Zemčík, Pavel (advisor)
Computer vision algorithms typically process real world image data acquired by cameras or video cameras. Such image data suffer from imperfections cause by the acquisition process. This paper focuses on simulation of the acquisition process on simulation of the acquisition process in order to enable rendering of images based on a 3D generated model. Imperfections, such as geometry distorion, chromatic aberration, depth of field effect, motion blur, vignetting and lens flare are considered.
|
|
|
Picture correction
Sedlo, Petr ; Horák, Karel (referee) ; Richter, Miloslav (advisor)
This work is focussed on different unwanted effects that damage a digitally acquired image. An image acquired using a digital camera can suffer from a series of defects. First, the noise is concerned. It occures due to discretisation. Next, there are the defects including vignetting, radial image distortion, chromatic aberration. In my thesis I also worked on backlight image compensation and panorama creation from an image sequence. For backlight images we need to increase the brightness in the dark parts of the image, so that the details became more visible. At panorama creation we are careful about brightness evenness in partial images, which is usually not conserved and consequently we try to find a suitable boundary-line to connect images and smooth it. The fundamental part of my work was writing the algorithms that are able to compensate these defects at least partly. For this aim I took use of the Matlab environment.
|
|
|
Simulation of Lens Features
Kučiš, Michal ; Seeman, Michal (referee) ; Zemčík, Pavel (advisor)
Computer vision algorithms typically process real world image data acquired by cameras or video cameras. Such image data suffer from imperfections cause by the acquisition process. This paper focuses on simulation of the acquisition process on simulation of the acquisition process in order to enable rendering of images based on a 3D generated model. Imperfections, such as geometry distorion, chromatic aberration, depth of field effect, motion blur, vignetting and lens flare are considered.
|
guest ::
