A toolbox for predicting perceived motion artifacts on modern displays

D. Galor*, G. Meng*, Laura Waller, Martin Banks

Published in Optica special issue on Imaging Systems, Microscopy, and Displays.

Dekel Galor*, Guanghan Meng*, Laura Waller, and Martin S. Banks, "BiPMAP: a toolbox for predicting perceived motion artifacts on modern displays," Opt. Express 32, 12181-12199 (2024) https://doi.org/10.1364/OE.510985

Overview

• Designed and implemented a toolbox for VR companies to optimize the tradeoff between display performance and cost.

• Includes a powerful DSP pipeline that simulates human visual perception.

• Using this pipeline, the tool predicts the perceived motion artifacts for a display.

Details

• Developed a DSP pipeline to model the human visual system.

• Created a novel formula for a human visual model (CSF formula).

• Created a software toolbox for simulating human perception during display design.

• Used Python, Flask backend, and React.js frontend for toolbox

• Deployed application with CUDA to accelerate computation

The project is supported by Berkeley CIVO. Special thanks to NVIDIA for providing us with compute resources.

Check out my amazing mentor Guanghan Meng. Also check out our lab websites for Prof. Laura Waller and for Prof. Martin Banks (although he is probably climbing a mountain at the moment).