Event camera system based on a liquid crystal microlens array for high dynamic range imaging in static scenes

Area of Science:

• Computer Vision
• Optical Engineering
• Image Processing

Background:

• Event cameras provide high temporal resolution but struggle with sparse data in static scenes, limiting high dynamic range (HDR) imaging.
• Existing event-based HDR methods face challenges in static environments due to limited event generation.

Purpose of the Study:

• To develop a novel liquid crystal modulated event camera (LC-MEC) system for enhanced HDR imaging in static scenes.
• To enable the construction of informative event streams from static scenes for improved HDR reconstruction.

Main Methods:

• An integrated LC-MEC system was designed, incorporating a liquid crystal microlens array (LC-MLA) to modulate the point spread function (PSF).
• Discrete voltages applied to the LC-MLA introduced intensity variations, enabling event stream construction from continuous intensity sequences.
• High dynamic range images were reconstructed by fusing intensity frames and event streams using a temporal event fusion network (TEF-Net).

Main Results:

• The LC-MEC system successfully produced structured and informative event streams even in static scenes.
• Significant improvements in the quality of high dynamic range image reconstruction were demonstrated.
• The proposed system effectively overcomes the limitations of sparse event generation in static environments.

Conclusions:

• The LC-MEC system offers a viable solution for HDR imaging in static scenes, overcoming the sparsity issue of traditional event cameras.
• The integration of LC-MLA modulation with deep learning-based fusion provides a powerful approach for advanced event-based imaging.
• This research advances the capabilities of event cameras for challenging imaging conditions.