Stereographic
Combing a Porcupine
or
Studies on
Orientation Diffusion
Nir A. Sochen, Chen Sagiv, Ron Kimmel
In this page the color figures appearing in the manuscript are
presented.
Figure 6.1
|
|
|
Two vector fields before (left) and
after (right) the
flow on S1.
|
Figure 6.2
|
|
|
The colors are restricted to one
quarter of the upper hemisphere defined around the black point in the RGB
space.
|
|
|
|
|
The original image (left), the
noisy image (middle), and the filtered image (right).
|
|
|
|
|
The vector fields of part of the
images, before (left) and after (right) the flow.
|
Figure 6.3
|
|
|
The HSV color model captures human
color perception better than the RGB model which is the common way our
machines represent colors.
|
|
|
|
|
The original image (left), the
noisy image (middle) and the filtered image (right) demonstrate the effect of
the flow as a denoising filter in the HSV color space.
|
Figure 6.4
|
|
|
An example for stereographic
orientation diffusion used in the HSV color space. The original image (left),
the noisy image (middle) and the filtered image (right) demonstrate the
effect of the flow as a denoising filter in the HSV color space when using
stereographic coordinates.
|
Figure 7.13
|
|
|
a. The original noise-free image
(left). b. The image after random noise was added (right).
|
Figure 7.14
|
|
|
a. The result of Linear diffusion
(left). b. The result of TV diffusion (right).
|
Figure 7.15
|
|
|
a. The result of HP diffusion for b = 10-5 (left). b. The result of HP
diffusion for b = 10 (right).
|
Figure 7.16
|
|
|
a. The result of SP diffusion for b = 10-3 (left). b. The result of SP
diffusion for b = 10 (right).
|
Last Modified: Thursday, 10 October 2002