search

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

How to select the theta_alpha, theta_beta and theta_gamma? #155

Open glhfgg1024 opened 6 years ago

glhfgg1024 commented 6 years ago

In CRFasRNN, the theta_alpha and the theta_gamma are for spatial terms, and the theta_beta is for the term in feature space. In https://github.com/torrvision/crfasrnn/blob/22db6d8a1a69ecafe97085d44a4d05eeec299634/python-scripts/TVG_CRFRNN_COCO_VOC.prototxt#L144, the theta_alpha is 160, but the theta_gamma is 3. Why did these two parameters set for so much difference? In addition, the theta_alpha is 160 means that the image has a big spatial blur, right? What does this mean? Does any know the reason? Thanks.