search

POSTECH-IMLAB / LaneSegmentationNetwork

The deep semantic segmentation network for lane segmentation.
GNU Lesser General Public License v3.0
95 stars 51 forks source link
deep-learning lane-detection lane-segmentation semantic-segmentation semantic-segmentation-network tensorflow tensorflow-semantic-segmentation

readme

LaneSegmentationNetwork

Created by Dongkyu Yu and Daijin Kim at POSTECH IM Lab

Demo

Youtube

Lane segmentation demo

Overview

The semantic segmentaion network for lane segmentation. Which inspired by Google DeepLabV3. We use the ResNet as backbone network for high quailty feature extraction. And we design the module which employ atrous convolution with multi atrous rate which use same filters. It makes not only the network robust to multiple scales but also reduce the number of parameters for filters.

Performance

We conduct the experiments on Highway Driving Dataset for Semantic Video Segmentation from KAIST, achieving the test set performance 87.6% mIoU.

Pre-trained model

You can download the pretrained model of network here.

Acknowledgements

This work was supported by Institute for Information & communications Technology Promotion (IITP) grant funded by the Korea government (MSIP)(2014-0-00059, Development of Predictive Visual Intelligence Technology), MSIP (Ministry of Science, ICT and Future Planning), Korea, under the ICT Consilience Creative Program (IITP-R0346-16-1007) supervised by the IITP, and MSIP(Ministry of Science, ICT and Future Planning), Korea, under the ITRC (Information Technology Research Center) support program (IITP-2017-2016-0-00464) supervised by the IITP.