Addition of 'Edge Drawing' and 'Edge Drawing Parameter Free' algorithms to skimage?

Robin Wilson r.t.wilson.bak at googlemail.com
Fri May 13 15:35:30 EDT 2016


Hi Egor,

The paper on the basic ED algorithm is available 
at https://www.researchgate.net/profile/Cuneyt_Akinlar/publication/257466884_Edge_Drawing_A_combined_real-time_edge_and_segment_detector/links/55daba4008ae9d659491f9c0.pdf

I can't find a freely available copy of the EDPF paper, but to be honest 
the descriptions on the websites I linked to are very similar to the paper 
content - exact copies in some cases!

I've actually had better performance than I was expecting, and even better 
performance when using numba for JIT compilation. I assume I'd get similar 
improvements with Cython.

Cheers,

Robin

On Friday, 13 May 2016 20:22:21 UTC+1, Egor Panfilov wrote:
>
> Hi Robin!
> That is very interesting algorithm! Unfortunately I couldn't find free 
> paper on it.
> I think there might be some performance issues for pure Python 
> implementation.
>
> Anyways, feel free to submit a PR (even if it is still work in process) 
> and I'll be happy to review it.
>
> Regards,
> Egor
>
> пятница, 13 мая 2016 г., 13:02:17 UTC+3 пользователь Robin Wilson написал:
>>
>> Hi,
>>
>> I'm currently in the process of implementing the Edge Drawing algorithm 
>> (see http://ceng.anadolu.edu.tr/CV/EdgeDrawing/) in Python, and am just 
>> starting to implement the additions required to produce the Edge Drawing 
>> Parameter Free (http://ceng.anadolu.edu.tr/CV/EDPF/) algorithm too.
>>
>> In summary, these are quite interesting and useful edge detection 
>> algorithms that use a 'joining the dots' approach, and so produce nicely 
>> connected edges. The Parameter Free version also seems to work remarkably 
>> well for an algorithm that doesn't require any parameters!
>>
>> Would the addition of these algorithms to skimage be of interest?
>>
>> The implementation isn't quite finished yet, and could do with some 
>> tidying (potentially with advice from some of the skimage maintainers), but 
>> it shouldn't be too long until they are in a fit state.
>>
>> Cheers,
>>
>> Robin
>>
>> Dr Robin Wilson
>> Research Fellow
>> University of Southampton, UK
>>
>>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/scikit-image/attachments/20160513/3f70413a/attachment.html>


More information about the scikit-image mailing list