Skip to content

Latest commit

 

History

History
15 lines (11 loc) · 682 Bytes

README.md

File metadata and controls

15 lines (11 loc) · 682 Bytes

ConvCRF

This a Tensorflow implementation of ConvCRF, which is proposed in paper "Convolutional CRFs for Semantic Segmentation" writed by Marvin T. T. Teichmann and Roberto Cipolla. This repository only contain the ConvCRF method , which is implemented as Layer class, so it is easy to connect to your model. The ConvCRF Layer is programmed by reference to ConvCRF pyTorch and crfasrnn_keras.

Requirements

Plattform: Linux, python2 >= 2.7, tensorflow-1.8, cuda 9.0, cudnn 7.0

Python Packages: numpy