Need help with npbg?
Click the “chat” button below for chat support from the developer who created it, or find similar developers for support.

About the developer

144 Stars 22 Forks MIT License 22 Commits 2 Opened issues


Neural Point-Based Graphics

Services available


Need anything else?

Contributors list

# 6,829
6 commits
# 626,301
1 commit

Neural Point-Based Graphics

Project Video Paper

Neural Point-Based Graphics
Kara-Ali Aliev1   Artem Sevastopolsky1,2   Maria Kolos1,2   Dmitry Ulyanov3   Victor Lempitsky1,2
1Samsung AI Center   1Skolkovo Institute of Science and Technology

UPD (09.02.2021): added a Docker container which can be executed on a headless node. See Readme.


This is PyTorch implementation of Neural Point-Based Graphics (NPBG), a new method for realtime photo-realistic rendering of real scenes. NPBG uses a raw point cloud as the geometric representation of a scene, and augments each point with a learnable neural descriptor that encodes local geometry and appearance. A deep rendering network is learned in parallel with the descriptors, so that new views of the scene can be obtained by passing the rasterizations of a point cloud from new viewpoints through this network.


The following instructions describe installation of conda environment. If you wish to setup the Docker environment, see the Readme in the docker folder. This way is also recommended for headless machines (without X server enabled).

Run this command to install python environment using conda:

source scripts/


You can render one of the fitted scenes we provide right away in the real-time viewer or fit your own scene.

Download fitted scenes and universal rendering network weights from here and unpack in the sources root directory.

We suppose that you have at least one GeForce GTX 1080 Ti for fitting and inference.

Viewer navigation:

  • Rotation: press left mouse button and drag
  • Move: press rigth mouse button and drug / scroll middle mouse botton
  • Pan: press middle mouse button and drug

Use fitted scene

Here we show a couple of examples how to run fitted scenes in the viewer.

Person 1

python --config downloads/person_1.yaml --viewport 2000,1328 --origin-view

Since this scene was fitted on 4k images, we crop image size with

argument to fit the scene into memory of a modest GPU.


python --config downloads/studio.yaml --rmode fly


directory for more examples.

Fit your scene

Fitting a new scene consists of two steps:

  1. Point cloud reconstruction
  2. Fitting descriptors

There is a bunch of software for point cloud reconstruction. While it is possible to adopt different software packages for our pipeline, we will choose Agisoft Metashape for this demonstration.

Point cloud reconstruction (Agisoft Metashape)

If you don't have a license for Agisoft Metashape Pro, start a trial version by filling in the form. On the first start, enter you license key.

Download and install Agisoft Metashape:

tar xvf metashape-pro_1_6_2_amd64.tar.gz
cd metashape-pro
LD_LIBRARY_PATH="python/lib:$LD_LIBRARY_PATH" ./python/bin/python3.5 -m pip install pillow

Optionally, enable GPU acceleration by checking Tools -> Preferences -> GPU.

Depending on the specs of your PC you may need to downscale your images to proper size. We recommend using 4k images or less. For example, if you want to downscale images by a factor of two, run this command: ```bash

convert comes with imagemagick package

sudo apt install imagemagick

in images directory

for fn in *jpg; do convert $fn -resize 50% $fn; done ```

Build point cloud:

bash -r /scripts/ 
 is the path to NPBG sources, 
 is directory with 
subdirectory with your scene images.

The script will produce: *

: dense point cloud *
: camera registration data *
: undistorted images for descriptor fitting *
: Metashape project *
: scene configuration for the NPBG viewer

Make sure the point cloud has no severe misalignments and crop out unnecessary geometry to optimize memory consumption. To edit a scene, open

in Metashape GUI and export modified point cloud (File -> Export -> Export Points). See Issues section for further recommendations.

Now we can fit descriptors for this scene.

Fitting descriptors


by setting absolute paths to scene configuration file, target images and, optionally, masks. Add other scenes to this file if needed.

Fit the scene:

python --config configs/train_example.yaml --pipeline npbg.pipelines.ogl.TexturePipeline --dataset_names 
 is the name of the scene in 
. Model checkpoints and Tensorboard logs will be stored in

The command above will finetune weights of the rendering network. This regime usually produces more appealing results. To freeze the rendering network, use option

. We provide pretrained weights for the rendering network on ScanNet and People dataset located in
. Set pretrained network using
option in

If you have masks for target images, use option '--use_masks'. Make sure masks align with target images.

When the model converge (usually 10 epochs is enough), run the scene in the viewer:

python --config .yaml --checkpoint data/logs//checkpoints/.pth --origin-view


is the scene configuration file created in the point cloud reconstruction stage,
is the path to descriptors checkpoint and
option automatically moves geometry origin to the world origin for convenient navigation. You can manually assign
field in
for arbitrary origin transformation (see
for example).

Guidelines for fitting novel scenes

Fitting novel scenes can sometimes be tricky, most often due to the preparation of camera poses that are provided in different ways by different sources, or sometimes because of the reconstruction issues (see below). We recommend checking out this and this issues for detailed explanations.

The most important insight is related to the configs structure. There is a system of 3 configs used in NPBG:

(there is another optional config -- inference config, which is essentially a scene config with

parameters: paths to the network weights checkpoint and a descriptors checkpoint, respectively)

To fit a new scene, one should a scene config

and a path config
by setting absolute paths to scene configuration file, target images, and other optional parameters, such as masks. Path config can contain paths to images of either 1 scene or several scenes, if needed. Examples of all configs of all types can be found in the repository.

Code logic and structure

Since our repository is based on a custom, specific framework, we leave the following diagram with the basic code logic. For those who wish to extend our code with additional features or try out related ideas (which we would highly appreciate), this diagram should help finding where the changes should be applied in the code. At the same time, various technical intricacies are not shown here for the sake of clarity.


  • Reconstruction failure and geometry misalignment. Taking photos for photogrammetry is the most critical part in the whole pipeline. Follow these recommendations to have a good reconstruction:
    • Set the highest quality in the camera settings;
    • Keep the object in focus, don't set f-stop too low;
    • Fix shutter speed, the faster the better, but don't underexpose the photo;
    • Photos must be bright enough, but don't overexpose the photo;
    • Keep ISO low enough as it may introduce noise;
    • Keep the objects still, remove moving parts from the scene;
    • Take photos with at least 70% overlap;
    • If possible, use soft diffused light;
    • Avoid dark shadows;

If you are using a smarthone with Android, OpenCamera may come handy. A good starting point for settings is f/8, ISO 300, shutter speed 1/125s. iPhone users are recommended to fix exposure in the Camera. Follow this guide for more recommendations. * Viewer performance. If Pytorch and X server run on different GPUs there will be extra data transfer overhead between two GPUs. If higher framerate is desirable, make sure they run on the same GPU (use

). * Pytorch crash on train. there is a known issue when Pytorch crashes on backward pass if there are different GPUs, f.e. GeForce GTX 1080 Ti and GeForce RTX 2080 Ti. Use
to mask GPU.


This is what we want to implement as well. We would also highly appreciate the help from the community.

  • Point cloud reconstruction with COLMAP. As Agisoft Metashape is a proprietary software, the community would most benefit from an open source package like COLMAP which has almost the same functionality as Metashape, so the goal is to have
    script working in the same manner as
  • Convenient viewer navigation. Interactively choosing rotation center would make navigation much more conveniet. At the moment the viewer either explicitly imports the model's origin matrix or sets the origin automatically based on point cloud density.


    title={Neural Point-Based Graphics},
    author={Kara-Ali Aliev and Artem Sevastopolsky and Maria Kolos and Dmitry Ulyanov and Victor Lempitsky},

We use cookies. If you continue to browse the site, you agree to the use of cookies. For more information on our use of cookies please see our Privacy Policy.