[Sidefx-houdini-list] Surface from image stack

Louis Dunlevy louisdunlevy at gmail.com
Thu Feb 11 18:10:07 EST 2010


geometry cop -> trace sop?

Mark Bolstad wrote:
> Posted to the forum:
>
> http://www.sidefx.com/index.php?option=com_forum&Itemid=172&page=viewtopic&p=86172#86172
> .
>
> Thanks,
>
> Mark
>
>
> 2010/2/11 François Duchesneau <sidefx at trinix.ca>
>
>   
>> You'll have to continue this thread on the forum to attach images.
>>
>> Mark Bolstad wrote:
>>     
>>> Thanks for the ideas.
>>>
>>> If the list will allow me to attach images, I'll append 3 images that are
>>> representative of consecutive levels. This one has a bifurcation on level
>>> two.
>>>
>>> Mark
>>>
>>>
>>> 2010/2/11 François Duchesneau <sidefx at trinix.ca>
>>>
>>>
>>>       
>>>> That reminds me a tool I created to mimick Blast Code. In Blast Code you
>>>> could use a bad quality crack pattern found on the web with varying line
>>>> thickness. The tool I created used the Trace Sop internally but had to
>>>> do lots of Sop operation to make it work. However, it was very picky
>>>> about the line width. I'd be curious to know what is the algorithm Blast
>>>> Code uses because it looks more like something giving the center of the
>>>> line only, like you want. That little improvement on the Trace Sop would
>>>> be great.
>>>>
>>>> François
>>>>
>>>> Michael K. O'Brien wrote:
>>>>
>>>>         
>>>>> Hola~
>>>>>
>>>>>
>>>>>
>>>>>           
>>>>>> -- On Thu, Feb 11, 2010 at 04:29:58PM -0500, Jerry wrote:
>>>>>> Check the second tab on the trace SOP.
>>>>>>
>>>>>>
>>>>>>             
>>>>> I tried that, but without any luck. I'll give it another go.
>>>>>
>>>>> My sample image is 16x16, all black, white horiz line in the middle. I
>>>>>
>>>>>           
>>>> end up
>>>>
>>>>         
>>>>> with with a closed curve that encircles all of the pixels.
>>>>>
>>>>> I ended up writing a medial axis node, which is really slow, but seems
>>>>>           
>> to
>>     
>>>> be a
>>>>
>>>>         
>>>>> bit closer to what I would expect.
>>>>>
>>>>> MO
>>>>>
>>>>>
>>>>>
>>>>>           
>>>>>>> Hola~
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>>               
>>>>>>>> -- On Thu, Feb 11, 2010 at 03:19:05PM -0500, Mark Elendt wrote:
>>>>>>>> On Thursday Feb 11 at 14:46, Mark Bolstad wrote:
>>>>>>>>
>>>>>>>>
>>>>>>>>                 
>>>>>>>>> I have a stack of black and white images that consist of a set of
>>>>>>>>> disconnected circle like shapes. First, how do you extract the
>>>>>>>>>
>>>>>>>>>
>>>>>>>>>                   
>>>>>>>> boundary
>>>>>>>>
>>>>>>>>
>>>>>>>>                 
>>>>>>>>> curve(s) and pass them to a SOP?
>>>>>>>>>
>>>>>>>>> Second, given a set of these extracted curves, I need to join them
>>>>>>>>>
>>>>>>>>>
>>>>>>>>>                   
>>>>>>>> into a
>>>>>>>>
>>>>>>>>
>>>>>>>>                 
>>>>>>>>> surface (assume some fixed offset between the curves). The problem
>>>>>>>>>                   
>> is
>>     
>>>>>>>>>                   
>>>>>>>> that
>>>>>>>>
>>>>>>>>
>>>>>>>>                 
>>>>>>>>> in some cases I'll have two curves on one level and 1 curve on the
>>>>>>>>>
>>>>>>>>>
>>>>>>>>>                   
>>>>>>>> next
>>>>>>>>
>>>>>>>>
>>>>>>>>                 
>>>>>>>>> level (the actual data is bifurcating somewhere in between the
>>>>>>>>>
>>>>>>>>>
>>>>>>>>>                   
>>>>>>>> levels).
>>>>>>>>
>>>>>>>>
>>>>>>>>                 
>>>>>>>>> The application is reconstructing neurons from a set of electron
>>>>>>>>>
>>>>>>>>>
>>>>>>>>>                   
>>>>>>>> microscopy
>>>>>>>>
>>>>>>>>
>>>>>>>>                 
>>>>>>>>> images.
>>>>>>>>>
>>>>>>>>> Any help is appreciated.
>>>>>>>>>
>>>>>>>>>
>>>>>>>>>                   
>>>>>>>> Could you make a volume (even an i3d file)?
>>>>>>>> The ISO SOP with an inside/out expression?
>>>>>>>>
>>>>>>>> If the topology changes (i.e. you have two circles blending to one),
>>>>>>>> that could be problematic.
>>>>>>>>
>>>>>>>> If you are guaranteed to have only one circle, then use the trace
>>>>>>>>                 
>> SOP
>>     
>>>>>>>> and poly loft (or poly spline and skin).
>>>>>>>>
>>>>>>>>
>>>>>>>>                 
>>>>>>> Is there a way to get a true curve out of trace? Trace wants to
>>>>>>>
>>>>>>>               
>>>> surround
>>>>
>>>>         
>>>>>>> each
>>>>>>> pixel with a poly face, as opposed to generating a curve that
>>>>>>>               
>> connects
>>     
>>>>>>> the
>>>>>>> pixels above the threshold, right?
>>>>>>>
>>>>>>> MO
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>>               
>>>>>>>> --
>>>>>>>> Mark Elendt
>>>>>>>> _______________________________________________
>>>>>>>> Sidefx-houdini-list mailing list
>>>>>>>> Sidefx-houdini-list at sidefx.com
>>>>>>>> https://lists.sidefx.com:443/mailman/listinfo/sidefx-houdini-list
>>>>>>>>
>>>>>>>>
>>>>>>>>                 
>>>>>> _______________________________________________
>>>>>> Sidefx-houdini-list mailing list
>>>>>> Sidefx-houdini-list at sidefx.com
>>>>>> https://lists.sidefx.com:443/mailman/listinfo/sidefx-houdini-list
>>>>>>
>>>>>>
>>>>>>             
>>>> _______________________________________________
>>>> Sidefx-houdini-list mailing list
>>>> Sidefx-houdini-list at sidefx.com
>>>> https://lists.sidefx.com:443/mailman/listinfo/sidefx-houdini-list
>>>>
>>>>
>>>> ------------------------------------------------------------------------
>>>>
>>>> _______________________________________________
>>>> Sidefx-houdini-list mailing list
>>>> Sidefx-houdini-list at sidefx.com
>>>> https://lists.sidefx.com:443/mailman/listinfo/sidefx-houdini-list
>>>>
>>>>         
>> _______________________________________________
>> Sidefx-houdini-list mailing list
>> Sidefx-houdini-list at sidefx.com
>> https://lists.sidefx.com:443/mailman/listinfo/sidefx-houdini-list
>>
>>     
> _______________________________________________
> Sidefx-houdini-list mailing list
> Sidefx-houdini-list at sidefx.com
> https://lists.sidefx.com:443/mailman/listinfo/sidefx-houdini-list
>
>   




More information about the Sidefx-houdini-list mailing list