This is the code for the paper STFCN: Spatio-Temporal FCN for Semantic Video Segmentation,
Mohsen Fayyaz, Mohammad Hajizadeh Saffar, Mohammad Sabokrou, Mahmood Fathy, Fay Huang, Reinhard Klette

A newer version of this work, "STFCN: Spatio-Temporal Fully Convolutional Neural Network for Semantic Segmentation of Street Scenes" has been accepted in ACCV 2016 Workshop on Computer Vision Technologies for Smart Vehicle.

Our Caffe prototxts are available. When we were working on this project, there was no official LSTM layer in Caffe. However, at the moment Caffe supports LSTM officially so we did not put our modified Caffe version (LSTM included) here.

The Torch source code for both locally and convolutional LSTM will be available as soon as possible...

If you found this work useful, please cite the following paper:

Mohsen Fayyaz, Mohammad Hajizadeh Saffar, Mohammad Sabokrou, Mahmood Fathy, Fay Huang, Reinhard Klette, “SFTCN: Spatio-Temporal Fully Convolutional Neural Network for Semantic Segmentation of Street Scenes,” 2016 Asian Conference on Computer Vision Workshop (ACCV Workshop), Taiwan, 2016, pp. 493-509. doi: 10.1007/978-3-319-54407-6_33