Re: [scikit-image] BPDL in skimage?
Hi Jiri On Mon, Oct 10, 2016, at 04:42, Jiří Borovec wrote:
recently we made a new method Binary Pattern Dictionary learning for estimation an atlas from set of binary images where we detect and decompose frequent patterns in the image set, see paper in attachment. I was wondering if you are interested to have this method as contribution to the scikit-image library? We typically wait a while for papers to gain popularity before we implement them in scikit-image.
That said, I think it's an excellent idea to make the code available with proper licensing and documentation until such time that we can include it. Best regards Stéfan
Dear All, Together with my colleagues I developed an approach for object recognition which we call COSFIRE. The approach was published on PAMI: G. Azzopardi and N. Petkov, “Trainable COSFIRE filters for keypoint detection and pattern recognition”, IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 35 (2), pp. 490-503, 2013. Since its publication we have used it in several applications with successful results. There is already the Matlab code available online: https://nl.mathworks.com/matlabcentral/fileexchange/37395-trainable-cosfire-... I thought it would be great if we could implement it in Python and release it for free in this community. What do you think? Best regards, George
Hi George On Tue, Oct 11, 2016, at 11:23, George Azzopardi wrote:
G. Azzopardi and N. Petkov, “Trainable COSFIRE filters for keypoint detection and pattern recognition”, IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 35 (2), pp. 490-503, 2013.
Since its publication we have used it in several applications with successful results. There is already the Matlab code available online: https://nl.mathworks.com/matlabcentral/fileexchange/37395-trainable-cosfire-...
I thought it would be great if we could implement it in Python and release it for free in this community. What do you think?
I would love to see a Python version out there, and to have a comparison done with the current keypoint detectors in scikit-image. With the steady increase in citations of your paper, and proof of good performance, I think this could be a good candidate for inclusion in scikit-image. Are you or one of your colleagues/students available to work on the implementation? Best regards Stéfan
Ok, it means wait until it will be generally proved that it is useful method, correct? -- Best regards, Jiri Borovec ------------------------------------------------------------------------ Ing. Jiri Borovec, MSc <jiri.borovec@fel.cvut.cz> PhD student at CMP CTU, http://cmp.felk.cvut.cz/~borovji3 On 11 October 2016 at 20:07, Stefan van der Walt <stefanv@berkeley.edu> wrote:
Hi Jiri
On Mon, Oct 10, 2016, at 04:42, Jiří Borovec wrote:
recently we made a new method Binary Pattern Dictionary learning for estimation an atlas from set of binary images where we detect and decompose frequent patterns in the image set, see paper in attachment. I was wondering if you are interested to have this method as contribution to the scikit-image library?
We typically wait a while for papers to gain popularity before we implement them in scikit-image.
That said, I think it's an excellent idea to make the code available with proper licensing and documentation until such time that we can include it.
Best regards Stéfan
participants (3)
-
George Azzopardi -
Jiří Borovec -
Stefan van der Walt