g******r
发帖数: 3 | 1 I am looking for papers and examples for volume segmentation. Bull people,
please, provide me the useful materials.
Thanks a lot. | l**p
发帖数: 474 | 2 What is your goal. Do some reserch or just need a simple solution
【在 g******r 的大作中提到】
: I am looking for papers and examples for volume segmentation. Bull people,
: please, provide me the useful materials.
: Thanks a lot.
| g******r
发帖数: 3 | 3 The volume segmentation is part of my current project. For a simple example,
I need to extract a human brain grey matter volume segment from the
3-D human brain gray-scale image. The process is done by highlighting a
volume of interested 2D region on the 3D gray-scale image, then the
corresponding 3D volume is extracted to a new window frame for viewing. In
addition, the extracted 3D volume is viewed as a color 3D image.
Anyone know certain papers(examples) related to my project, please,
【在 l**p 的大作中提到】
: What is your goal. Do some reserch or just need a simple solution
| l**p
发帖数: 474 | 4 .. for your way, you may use some AI classification method.
the selected region is the training dataset... there shoudl be many existing
ways to do the classification.
Kuan-Liu Ma has one paper in this year's Vis. Although that paper is mainly
about interface, you should be bale to find some refs there
example,
【在 g******r 的大作中提到】
: The volume segmentation is part of my current project. For a simple example,
: I need to extract a human brain grey matter volume segment from the
: 3-D human brain gray-scale image. The process is done by highlighting a
: volume of interested 2D region on the 3D gray-scale image, then the
: corresponding 3D volume is extracted to a new window frame for viewing. In
: addition, the extracted 3D volume is viewed as a color 3D image.
: Anyone know certain papers(examples) related to my project, please,
| e****r
发帖数: 166 | 5 just a thought(ignore me if you think it's nonsense),
if you can bind some depth information with your 2d image pixel, say
a z-buffer value,maybe it's more easy to extract from 3d image data set
after you hightlight the 2d region. is that simple enough?:o)
In
it.
【在 l**p 的大作中提到】
: .. for your way, you may use some AI classification method.
: the selected region is the training dataset... there shoudl be many existing
: ways to do the classification.
: Kuan-Liu Ma has one paper in this year's Vis. Although that paper is mainly
: about interface, you should be bale to find some refs there
:
: example,
| l**p
发帖数: 474 | 6 for the volume rendering, it might be difficult to get the depth information.
All non-empty voxel will effect the z-buffer. but it would be interest if you
can find someway to get around it.... make soem information you can direct
index to the interested part of the volume.
existing
mainly
a
【在 e****r 的大作中提到】
: just a thought(ignore me if you think it's nonsense),
: if you can bind some depth information with your 2d image pixel, say
: a z-buffer value,maybe it's more easy to extract from 3d image data set
: after you hightlight the 2d region. is that simple enough?:o)
:
: In
: it.
| e****r
发帖数: 166 | 7 yeah, i know what you mean, maybe the traditional z-buffer can't
handle such problem, but if we can us some kinda of 'volume'z-buffer,
say every non-empty voxel will have an index in this 'volume'z-buffer,
maybe we can archieve the goal. of course it's not efficient way, but
it might be good for realtime stuff after the preprocess.
information.
you
highlighting
viewing.
post
Bull
【在 l**p 的大作中提到】
: for the volume rendering, it might be difficult to get the depth information.
: All non-empty voxel will effect the z-buffer. but it would be interest if you
: can find someway to get around it.... make soem information you can direct
: index to the interested part of the volume.
:
: existing
: mainly
: a
| l**p
发帖数: 474 | 8 could be implemented in software.
and Ester, any progress in your own volume work?
【在 e****r 的大作中提到】
: yeah, i know what you mean, maybe the traditional z-buffer can't
: handle such problem, but if we can us some kinda of 'volume'z-buffer,
: say every non-empty voxel will have an index in this 'volume'z-buffer,
: maybe we can archieve the goal. of course it's not efficient way, but
: it might be good for realtime stuff after the preprocess.
:
: information.
: you
: highlighting
: viewing.
| e****r
发帖数: 166 | 9 haven't got time to improve it yet, i'm working my ass off on
two course projects now, but i will make some imporvement later on,
maybe use it to render my project result, two thing i really want
to add are isosurface and self shadowing, did you touch levelset
before, i think it's better than marching cube when tracking moving
volume surface.
if
direct
set
simple
the
the
please,
【在 l**p 的大作中提到】
: could be implemented in software.
: and Ester, any progress in your own volume work?
| l**p
发帖数: 474 | 10 Levelset is interesting. Do you have any new idea?
many
is
【在 e****r 的大作中提到】
: haven't got time to improve it yet, i'm working my ass off on
: two course projects now, but i will make some imporvement later on,
: maybe use it to render my project result, two thing i really want
: to add are isosurface and self shadowing, did you touch levelset
: before, i think it's better than marching cube when tracking moving
: volume surface.
:
: if
: direct
: set
| w**f
发帖数: 7794 | 11 levelset is interesting, but way hard to implement,
for marching cube, its very easy if you use classes in
VTK.
【在 l**p 的大作中提到】
: Levelset is interesting. Do you have any new idea?
:
: many
: is
| w**f
发帖数: 7794 | 12 still not clear how you highlight 2D ROI,
you mean on each single slice or on the 3D image?
【在 g******r 的大作中提到】
: The volume segmentation is part of my current project. For a simple example,
: I need to extract a human brain grey matter volume segment from the
: 3-D human brain gray-scale image. The process is done by highlighting a
: volume of interested 2D region on the 3D gray-scale image, then the
: corresponding 3D volume is extracted to a new window frame for viewing. In
: addition, the extracted 3D volume is viewed as a color 3D image.
: Anyone know certain papers(examples) related to my project, please,
| e****r
发帖数: 166 | 13 yes, there are a lot realtime blobby programs around, all you
need to do is copying the predifined huge data table and indexing it.
【在 w**f 的大作中提到】
: levelset is interesting, but way hard to implement,
: for marching cube, its very easy if you use classes in
: VTK.
|
|
