torrvision / crfasrnn

This repository contains the source code for the semantic image segmentation method described in the ICCV 2015 paper: Conditional Random Fields as Recurrent Neural Networks. http://crfasrnn.torr.vision/
Other
1.34k stars 462 forks source link

CRFasRNN Training #95

Open AlexTS1980 opened 7 years ago

AlexTS1980 commented 7 years ago

How did you train CRFasRNN, specifially what did the last layer look like?

What I mean is that it couldn't be just a vector of 21 neurons, because you have to segment the object in addition to classification. Did you use a mask size the input image with 1's in the location of the object and 0 elsewhere? I see that the output is of shape (batch_size, num_class, width, height), but not sure how you used this for training.