a module for 3D semantic segmentation in point clouds.
Switch branches/tags
Nothing to show
Clone or download
Latest commit cc18c4c Jul 14, 2018
Failed to load latest commit information.
img init Jul 1, 2018
models init Jul 1, 2018
tf_utils modify main.cpp Jul 14, 2018
README.md edit readme Jul 12, 2018
scannet_dataset.py init Jul 1, 2018
test.py init Jul 1, 2018
train_and_eval_scannet.py init Jul 1, 2018


PointSIFT: A SIFT-like Network Module for 3D Point Cloud Semantic Segmentation

Created by Mingyang Jiang, Yiran Wu, Cewu Lu (corresponding author).


PointSIFT is a semantic segmentation framework for 3D point clouds. It is based on a simple module which extract featrues from neighbor points in eight directions. For more details, please refer to our arxiv paper.

PointSIFT is freely available for free non-commercial use, and may be redistributed under these conditions. For commercial queries, contact Cewu Lu.

Fig1 not found!!


In our experiment, All the codes are tested in Python3.5(If you use Python 2.7, please add some system paths), CUDA 8.0 and CUDNN 5.1.

  1. Install TensorFlow (We use v1.4.1).
  2. Install other python libraries like h5py
  3. Compile TF operator (Similar to PointNet++). Firstly, you should find Tensorflow include path and library paths.
    import tensorflow as tf
    # include path
    # library path 

Then, change the path in all the complie file, like tf_utils/tf_ops/sampling/tf_sampling_compile.sh Finally, compile the source file, we use tf_sampling as example.

    cd tf_utils/tf_ops/sampling
    chmod +x tf_sampling_compile.sh


If you want use our model in your own project. After compiling the TF operator, you can import it easily. Here shows a simple case.(we take batch_size * num_point * input_dim as input and get batch_size * num_point * output_dim as output)

import tensorflow as tf
# import our module
from tf_utils.pointSIFT_util import pointSIFT_module
# input coordinates
xyz = tf.tf.placeholder(tf.float32, shape=(batch_size, num_point, 3))
# input features
point_feature = tf.tf.placeholder(tf.float32, shape=(batch_size, num_point, input_dim)
# setting phases
is_training = tf.placeholder(dtype=tf.bool, shape=())
# setting searching radius (0.1 as an example)
radius = 0.1
_, out_feature, _ = pointSIFT_module(xyz, point_feature, radius, output_dim, is_training)

Training and evaluating on ScanNet

  1. All the data can be download from here. They are the same as PointNet++.
  2. Train the data:
python train_and_eval_scannet.py

If you have multiple GPU:

CUDA_VISIBLE_DEVICES=0,1,2,3 python train_and_eval_scannet.py --gpu_num=4


Please cite the paper in your publications if it helps your research:

Author = {Mingyang Jiang and Yiran Wu and Cewu Lu},
Title = {PointSIFT: A SIFT-like Network Module for 3D Point Cloud Semantic Segmentation},
Year = {2018},
Eprint = {arXiv:1807.00652},