[Insight-users] an error when running ShapeDetection

[zhao]

Hi Luis,
 	
	OK now. Thanks a lot.
	But it seems that the extracted shape of ventricles is apparently narrower than the visual one in the original image, Which is true even after I greatly increased the "t" parameter. 
	Other segmentation results derived from different seed points also seem mismatch with original anatomy structures.

Zhao
   
>Hi Zhao,
>
>I tested the ShapeDetection example with
>the image input file that you sent.
>
>The image can be read without problem.
>
>What is faililng in your test is the
>process of writing the PGM files.
>
>The ShapeDetection program ask you for
>a directory where to put the pgm files.
>
>You are answering "pgm" which requires
>you to have previously created a
>subdirectory "pgm".
>
>If you have not created this subdirectory,
>the program will not  be able to write on
>it. (the program will not try to create
>the directory).
>
>Please try by creating a subdirectory
>(e.g. "pgm", or any other name) and providing
>this name to the Segmentation example.
>
>You can also try by just giving "." (e.g. the
>current directory) in which case the pgm
>files will be writen in the current directory.
>
>---
>
>An easy way of testing this example is to recompile
>it for "unsigned char" images. This is done by
>replacing line 36 in ShapeDetectApp.h
>
>The current line is:
>
>   typedef signed short InputPixelType;
>
>you can replace it by:
>
>   typedef unsigned char InputPixelType;
>
>
>With this pixel type you can use the BrainWeb
>images available at the ftp server :
>
>ftp://public.kitware.com/pub/itk/Data/BrainWeb
>
>You can use the attached parameter file:
>"parameters.txt" like:
>
>    ShapeDetection < parameters.txt
>
>It is using the brainweb image brainweb165a10f17.raw
>and using a seed point at voxe l78 107 98 which
>correspond to one of the ventricles.  A time =200.0
>is enough to segment the ventricles in this case.
>
>The whole process (including reading and writing
>images) takes about 193 seconds in a Pentium 3
>1 Ghz.
>
>
>Please let us know if you encounter further problems,
>
>
>Thanks
>
>
>  Luis
>
>
>==============================================
>
>zhao wrote:
>
>> Hi Luis,
>> 
>> 	This raw file is in the attachment.
>> 
>> 	In fact it is a 512 * 512 2-D raw file. I use it as a 128*128*8(2 Byte/pixel) file, just want to test the example.
>> 
>> 	Thanks!
>> 
>>  
>> 
>>>Hi Zhao,
>>>
>>>Could you please post the input file
>>>that you used for running the example ?
>>>
>>>
>>>The problem could be related to permissions
>>>for writing in a directory. It could also be
>>>that the input path doesn't actually correspond
>>>to the locations where the input files are to be
>>>found...
>>>
>>>Please let us know the content of the input file,
>>>
>>>Thanks
>>>
>>>
>>>Luis
>>>
>>>=====================================
>>>
>>>zhao wrote:
>>>
>>>
>>>>Hi all,
>>>>	I encountered an error when running ShapeDetection:
>>>>	
>>>>	Input file name: e:\data\avf.raw
>>>>	Input image big endian? [y|n]: n
>>>>	Input image size: 128 128 8
>>>>	PGM output directory: pgm
>>>>	Writing PGM files of the input volume.
>>>>	Caught an exception:
>>>>	Execute()
>>>>	Can't open file stream.
>>>>
>>>>	Could you give me some suggestion to sovle it ? thanks!
>>>>
>>>>Zhao ChenGuang
>>>>P.O.Box:010,
>>>>Dept. BME,
>>>>Shanghai Jiao Tong University,
>>>>1954# Hua Shan Road,
>>>>Shanghai,P.R.China,
>>>>200030
>>>>
>>>>