[GRASS-dev] Applying for GSoC - Image Segmentation

[Moritz Lennert]

On 04/04/13 05:59, Markus Neteler wrote:
> On Thu, Mar 28, 2013 at 12:57 PM, sandeep kumar
> <kumarsandeep2357 at gmail.com>  wrote:
>>
>> Hi,
>>
>>      My name is sandeep and i am doing my MS by Research in Spatial
>> Informatics at IIIT-Hyderabad. I am a good programmer and a quick learner.
>> This is the first time i am applying for GSoC.
>>
>>       Actually, my MS work is OBIA (Object Based Image Analysis) with
>> watershed segmentation. Though the work is mainly with watershed, i learned
>> about several other segmentation algorithms(Region growing,mean shift etc.).
>>
>>       I went through the GSoC-2013 ideas page and thought that the image
>> segmentation idea is quite apt for me to work on. That way i can participate
>> in GSoC and contribute to the software i daily use.
>>
>>      Kindly let me know if someone is working on this idea or can i have a
>> chance to work on this idea.
>
> There is an existing algorithm from the last year's GSoC:
> http://grass.osgeo.org/grass70/manuals/i.segment.html
>
> but it could be further enhanced as outlined in
> http://grasswiki.osgeo.org/wiki/GRASS_SoC_Ideas_2013#Imagery

Yes, I think it is important to note that what we are looking for mostly 
at this stage is to use the results of segmentation for object-based 
classification. A possible option might be to implement more 
segmentation algorithms (at the moment we only have region growing), but 
in my eyes this is less of a priority.

Moritz