PPEDNet is an encoder-decoder netwrok based on VGG16 framework for image semantic segmentation. This work focuses on the tradeoff between performance and speed.
Supported OS: the source code was tested on 64-bit Ubuntu 14.04.3 LTS Linux OS, and it should also be executable in other linux distributions.
Installation: Please compile the modified Caffe framework caffe-segnet (It supports all all necessaty layers for PPEDNet).
Pre-trained Model: Please download the VGG16 pre-traiend model (https://github.com/BVLC/caffe/wiki/Model-Zoo).
If you find PPEDNet useful in your research, please consider to cite our paper:
@inproceedings{ICIG2017PPEDNet,
title={PPEDNet: Pyramid Pooling Encoder-Decoder Network for Real-Time Semantic Segmentation},
author={Zhentao Tan, Bin Liu, and Nenghai Yu},
booktitle={ICIG},
year={2017}
}