What is it all about?
This project was created as a part of doctorate research by Oleksii Doronin.
It contains a modelling of the most common types of light field dipslays, including some algorithms to enhance the quality of their visual output without changing the hardware configuration.
What is light field display?
Light field (LF) display is a 3D display which is able to show different picture from different viewing angles, without the need for user to wear a special glasses or enabling position tracking.
The most common types of LF displays are lenslet-based (integral imaging, also most of multi-view, and more simple parallax barrier), projector-based (e.g., Holografika), and multiplication-based (e.g., MIT Tensor).
Where is the source code from your paper?
If you learned about this project from one of my papers, you can go to Published papers section.
This project was created and maintained on Windows 10 machine, but potentially can be easily transfered to another platform.
It has the following dependencies:
- OpenCV (tested 4.3) — mandatory.
- LFRayTracerPBRT — mandatory (for now).
You should build the project as follows.
- Make sure you have OpenCV build, preferably 4.3. If not, download it.
- Clone this project to some folder on your machine. E.g., in
D:/Work/
callgit clone https://github.com/LeksiDor/LFDisplay.git
. - Clone LFRayTracerPBRT to the same folder. E.g., in
D:/Work/
callgit clone --recursive https://github.com/LeksiDor/LFRayTracerPBRT.git
. - Build LFRayTracerPBRT. For detailed instructions, go to its project page. Shortly, you should execute a regular CMake build, while choosing build folder as, e.g.,
D:/Work/LFRayTracerPBRT/build
. Don't forget to checkPBRT_FLOAT_AS_DOUBLE
flag. Then, compile in both Debug and Release. - Call
Initialize3rdparty.sh
script fromD:/Work/LFDisplay/
folder. This script will instantiate the3rdparty
folder, which contains all necessary things from LFRayTracerPBRT. - Build this project using CMake. Choose build folder as
D:/Work/LFDisplay/build
. - Compile this project. If you have errors, it's not ok.
Oleksii Doronin, Erdem Sahin, Robert Bregovic, Atanas Gotchev
A Framework for Assessing Rendering Techniques for Near-Eye Integral Imaging Displays.
To appear in ICIP 2020 proceedings.
Source code folder: /src/ExampleICIP2020/
Source code Wiki: ExampleICIP2020
Oleksii Doronin, Robert Bregovic, Atanas Gotchev
Optimized 3D Scene Rendering on Projection-Based 3D Displays.
To appear in EUSIPCO 2020 proceedings.
Source code folder: /src/ExampleEUSIPCO2020/
Release: Windows_x64