Sorry I answered to Hristo without reading the all conversations. I agree with Jordi but the documentation is quite confusing. It is not easy to understand (without looking at the code) that this parameter implicitly set the number of layers...
Moreover, looking at the code I see also that the number of neurons in the input layer is equal to the number of bands in the input image. It's not really documented and worst I did not find any check about this condition in the class... I'll fill a bug about this. Manuel Le 16/09/2015 13:52, Jordi Inglada a écrit : > Hi, > > I think that I don't agree with Guillaume, unless I have misunderstood > the question (or maybe this is a matter of OTB version; I am using the > development version). > > The command line application allows to choose the number of neurons in > each hiddent layer with the following option: > > -classifier.ann.sizes <string list> Number of neurons in each > intermediate layer (mandatory) > > So if you use > > -classifier.ann.sizes 5 4 5 > > you will get 3 hidden layers with 5, 4 and 5 neurons in respectively > going from input towards output. > > In the GUI version of the application, if you choose the ANN classifier, > you can add layers with the "+" sign and set the number of neurons for > each layer. > > Jordi > > Guillaume Pasero <[email protected]> wrote: >> >> Hi, >> >> Using more than 1 hidden layer is supported by OpenCV, but not in the >> OTB wrapper class. You will have to develop a different machine learning >> model, based on the existing one : otb::NeuralNetworkMachineLearningModel >> >> Regards, >> Guillaume >> >> On 09/15/2015 03:55 PM, Hristo Nikolov wrote: >>> Dear all, >>> >>> I try to use OTB inside QGIS for ANN data classification. >>> I realized that in the GUI the only option is to use ANN >>> with one hidden layer. I'd like to have an experiment >>> with two hidden layers. >>> >>> Is it possible? >>> >>> Thank you. >>> Hristo >>> >>> -- >>> -- >>> Check the OTB FAQ at >>> http://www.orfeo-toolbox.org/FAQ.html >>> >>> You received this message because you are subscribed to the Google >>> Groups "otb-users" group. >>> To post to this group, send email to [email protected] >>> To unsubscribe from this group, send email to >>> [email protected] >>> For more options, visit this group at >>> http://groups.google.com/group/otb-users?hl=en >>> --- >>> You received this message because you are subscribed to the Google >>> Groups "otb-users" group. >>> To unsubscribe from this group and stop receiving emails from it, send >>> an email to [email protected] >>> <mailto:[email protected]>. >>> For more options, visit https://groups.google.com/d/optout. >> >> -- >> <www.c-s.fr> *Guillaume PASERO* >> Ingénieur d'études et développement >> *Business Unit E-SPACE & Geo Information* >> <https://thor.si.c-s.fr/blogs/cs-blogs-business/>*- Département >> APPLICATIONS* >> >> *CS Systèmes d'Information* >> Parc de la Grande Plaine - 5, Rue Brindejonc des Moulinais - BP 15872 >> 31506 Toulouse Cedex 05 - FRANCE >> +33 561 17 64 21 - [email protected] >> >> -- > -- Manuel GRIZONNET -- -- Check the OTB FAQ at http://www.orfeo-toolbox.org/FAQ.html You received this message because you are subscribed to the Google Groups "otb-users" group. To post to this group, send email to [email protected] To unsubscribe from this group, send email to [email protected] For more options, visit this group at http://groups.google.com/group/otb-users?hl=en --- You received this message because you are subscribed to the Google Groups "otb-users" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. For more options, visit https://groups.google.com/d/optout.
