Denoising events camera data by overlapping multiple encoded channels

Area of Science:

• Computer Vision
• Sensor Technology
• Signal Processing

Background:

• Event cameras capture dynamic scenes with high temporal resolution.
• Noise in event data can degrade performance in downstream applications.
• Current noise reduction methods often rely solely on algorithmic approaches.

Purpose of the Study:

• To introduce an opto-algorithmic method for effective noise removal in event camera data.
• To distinguish between genuine signal events and uncorrelated noise.
• To demonstrate the method's robustness and improved performance over algorithmic techniques.

Main Methods:

• An opto-algorithmic approach was developed for noise reduction.
• The optical image was split into two spatially encoding channels.
• These channels were overlapped on the event camera sensor to identify co-occurring events.

Main Results:

• The method successfully separated co-occurring signal events from noise events.
• The opto-algorithmic technique demonstrated improved performance compared to purely algorithmic methods.
• The approach is agnostic to various types of background noise.

Conclusions:

• The proposed opto-algorithmic method offers an effective solution for noise reduction in event camera data.
• This technique enhances the quality of event data by distinguishing signal from noise.
• The method shows promise for improving the reliability of event camera systems.