Event Camera Simulator Improvements via Characterized Parameters

Abstract

It has been more than two decades since the first neuromorphic Dynamic Vision Sensor (DVS) sensor was invented, and many subsequent prototypes have been built with a wide spectrum of applications in mind. Competing against state-of-the-art neural networks in terms of accuracy is difficult, although there are clear opportunities to outperform conventional approaches in terms of power consumption and processing speed. As neuromorphic sensors generate sparse data at the focal plane itself, they are inherently energy-efficient, data-driven, and fast. In this work, we present an extended DVS pixel simulator for neuromorphic benchmarks which simplifies the latency and the noise models. In addition, to more closely model the behaviour of a real pixel, the readout circuitry is modelled, as this can strongly affect the time precision of events in complex scenes. Using a dynamic variant of the MNIST dataset as a benchmarking task, we use this simulator to explore how the latency of the sensor allows it to outperform conventional sensors in terms of sensing speed.

Real-Time Event-Based Unsupervised Feature Consolidation and Tracking for Space Situational Awareness

“…Similar to many conventional tracking papers, we use simulated tracking data to separate the accuracy and performance of a sensor from the tracking algorithm. We simulate targets observed using the sensor with the implementation described in Joubert et al (2021), whose radiometric inputs are estimated based distortion-free point spread functions matching real-world observations of RSO at varying brightness with our telescopes. With this method, we create simulated leap-frogs of single targets transiting through the simulated FOV at random angles 1 to separate the capabilities of the EBC and telescope setup from the performance of the proposed FIESTA algorithm.…”

Research of event cameras in space object detection

“…The V2e tool uses a non-ideal DVS pixel model to generate events from videos and extends the model to low lighting conditions [40] . The IEBCS simulator has extended ESIM and V2e, optimized the neural morphology benchmark of the model, significantly improved noise simulation, and is closer to the real pixel working mode [41] . Subsequently, the event camera simulator was applied to research on space object detection.…”

Neuromophic High-Frequency 3d Dancing Pose Estimation in Dynamic Environment

“…As for simulators, although there are some existing event camera simulators [40,45,60,66], there is a common and crucial problem. Almost all of them only work to turn existing pictures or videos into event streams instead of making highly customized event streams directly based on the research problem needs.…”