On the Theory of Bias Tuning in Event Cameras

David El-Chai Ben-Ezra; Daniel Brisk; Adar Tal

cs.CVJan 30, 2025

On the Theory of Bias Tuning in Event Cameras

Authors: David El-Chai Ben-Ezra, Daniel Brisk, Adar Tal

Abstract

This paper lays the foundation of a theory for bias tuning in neuromorphic cameras, a novel sensing technology also known as "event cameras". We begin by formulating the high-level effect of the sensitivity biases on the camera's event rate in mathematical terms. We then show that, as a corollary of the Poincare-Miranda theorem, the commonly used tuning principles of rate budgeting and polarity balancing lead to a unique configuration of the sensitivity biases. As a corollary, we show how by adopting these principles, the multi-variable bias-tuning problem reduces to a two-parameter problem that can be resolved experimentally.

Figures & tables

Figure 3.1: Validation of Equations 3.1 and 3.2 in a representative scenario. The event rates

R_{P}

and

R_{N}

are shown as functions of the sensitivity biases, with all other camera biases fixed at their default values. The left panel presents a sweep of

p

, while the right panel presents a sweep of

n

. The orange and green curves depict the positive and negative event rates, respectively. The red and blue curves show the corresponding seven-sample moving geometric means, highlighting the underlying trends in the positive and negative event-rate measurements.

Figure 3.2: Validation of the assumptions in Equation 3.3 using the smoothed event-rate curves shown in Figure 3.1 .

Figure 5.1: Values of the sensitivity biases $p$ = bias_diff_on and $n$ = bias_diff_off as functions of the filter biases $l$ = bias_fo and $h$ = bias_hpf. For each sampled $(l,h)$ , the sensitivity biases were tuned to maintain a polarity balanced event rate of approximately 100k events/sec.

Figure 5.2: Number of positive and negative events per period of the signal generated by the incandescent lamp, as a function of the filter biases.

Event-Based Vision

Event-based vision uses asynchronous camera measurements to perceive motion and scene changes with high temporal resolution and dynamic range.

See all 227 papers

BiasBench: A reproducible benchmark for tuning the biases of event cameras
Apr 25, 2025Andreas Ziegler, David Joseph, Thomas Gossard +2Event-Based Vision Benchmark Design
Cognitive Systems Group, University of Tübingen, Germany
Noise2Params: Unification and Parameter Determination from Noise via a Probabilistic Event Camera Model
May 4, 2026Owen Root, Julinda Mujo, Min XuEvent-Based Vision Parameter Estimation
Department of Physics and Astronomy, CUNY Hunter College, 695 Park Ave, New York, 10065, NY, USA.
Sensor Generalization for Adaptive Sensing in Event-based Object Detection via Joint Distribution Training
Feb 26, 2026Aheli Saha, René Schuster, Didier StrickerAdaptive Sensing Event-Based Vision
Deutsches Forschungszentrum für Künstliche Intelligenz, Kaiserslautern, Germany

Apr 25, 2025J/K move · Enter open · S save

BiasBench: A reproducible benchmark for tuning the biases of event cameras

Andreas Ziegler, David Joseph, Thomas Gossard, Emil Moldovan+1

Cognitive Systems Group, University of Tübingen, Germany

Event-Based Vision Benchmark Design

Open paper AskPDF arXiv

Event-based cameras are bio-inspired sensors that detect light changes asynchronously for each pixel. They are increasingly used in fields like computer vision and robotics because of several advantages over traditional frame-based cameras, such as high temporal resolution, low latency, and high dynamic range. As with any camera, the output's quality depends on how well the camera's settings, called biases for event-based cameras, are configured. While frame-based cameras have advanced automatic configuration algorithms, there are very few such tools for tuning these biases. A systematic testing framework would require observing the same scene with different biases, which is tricky since event cameras only generate events when there is movement. Event simulators exist, but since biases heavily depend on the electrical circuit and the pixel design, available simulators are not well suited for bias tuning. To allow reproducibility, we present BiasBench, a novel event dataset containing multiple scenes with settings sampled in a grid-like pattern. We present three different scenes, each with a quality metric of the downstream application. Additionally, we present a novel, RL-based method to facilitate online bias adjustments.

Abstract

Event-Based Vision

Explore similar work

BiasBench: A reproducible benchmark for tuning the biases of event cameras

Noise2Params: Unification and Parameter Determination from Noise via a Probabilistic Event Camera Model

Sensor Generalization for Adaptive Sensing in Event-based Object Detection via Joint Distribution Training