Implementation of SqueezeSeg, convolutional neural networks for LiDAR point clout segmentation
Clone or download
Pull request Compare This branch is even with BichenWuUCB:master.
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Type Name Latest commit message Commit time
Failed to load latest commit information.

SqueezeSeg: Convolutional Neural Nets with Recurrent CRF for Real-Time Road-Object Segmentation from 3D LiDAR Point Cloud

By Bichen Wu, Alvin Wan, Xiangyu Yue, Kurt Keutzer (UC Berkeley)

This repository contains a tensorflow implementation of SqueezeSeg, a convolutional neural network model for LiDAR segmentation. A demonstration of SqueezeSeg can be found below:

Please refer to our video for a high level introduction of this work: For more details, please refer to our paper: If you find this work useful for your research, please consider citing:

    title={Squeezeseg: Convolutional neural nets with recurrent crf for real-time road-object segmentation from 3d lidar point cloud},
    author={Wu, Bichen and Wan, Alvin and Yue, Xiangyu and Keutzer, Kurt},
    journal={arXiv preprint arXiv:1710.07368},


SqueezeSeg is released under the BSD license (See LICENSE for details). The dataset used for training, evaluation, and demostration of SqueezeSeg is modified from KITTI raw dataset. For your convenience, we provide links to download the converted dataset, which is distrubited under the Creative Commons Attribution-NonCommercial-ShareAlike 3.0 License.


The instructions are tested on Ubuntu 16.04 with python 2.7 and tensorflow 1.0 with GPU support.

  • Clone the SqueezeSeg repository:

    git clone

    We name the root directory as $SQSG_ROOT.

  • Setup virtual environment:

    1. By default we use Python2.7. Create the virtual environment

      virtualenv env
    2. Activate the virtual environment

      source env/bin/activate
  • Use pip to install required Python packages:

    pip install -r requirements.txt


  • To run the demo script:
    cd $SQSG_ROOT/
    python ./src/
    If the installation is correct, the detector should write the detection results as well as 2D label maps to $SQSG_ROOT/data/samples_out. Here are examples of the output label map overlaped with the projected LiDAR signal. Green masks indicate clusters corresponding to cars and blue masks indicate cyclists.


  • First, download training and validation data (3.9 GB) from this link. This dataset contains LiDAR point-cloud projected to a 2D spherical surface. Refer to our paper for details of the data conversion procedure. This dataset is converted from KITTI raw dataset and is distrubited under the Creative Commons Attribution-NonCommercial-ShareAlike 3.0 License.

    cd $SQSG_ROOT/data/
    tar -xzvf lidar_2d.tgz
    rm lidar_2d.tgz
  • Now we can start training by

    cd $SQSG_ROOT/
    ./scripts/ -gpu 0 -image_set train -log_dir ./log/

    Training logs and model checkpoints will be saved in the log directory.

  • We can launch evaluation script simutaneously with training

    cd $SQSG_ROOT/
    ./scripts/ -gpu 1 -image_set val -log_dir ./log/
  • We can monitor the training process using tensorboard.

    tensorboard --logdir=$SQSG_ROOT/log/

    Tensorboard displays information such as training loss, evaluation accuracy, visualization of detection results in the training process, which are helpful for debugging and tunning models, as shown below: alt text alt text