Stefan Isler

Efficiency Through Adaptive Sensing Using Event-based Sensors

Abstract: To appropriately address the complex, challenging and diverse requirements of modern Computer Vision in the robotics, consumer electronics, automotive or automation sectors Sony AVS takes a holistic sensing system design approach. In our developments we carefully weigh strengths and weaknesses of both sensing modalities and processing hardware to complement each other. I will talk about how we combine the strength of Sony’s sensors such as conventional frame-based cameras, depth sensors and event-based sensors (EVS) towards building optimal visual sensing systems. In particular, we utilize the high temporal resolution and spatial sparsity of event-based sensor data to both increase the efficiency and reduce the latency in our vision stack through an adaptive processing paradigm

Stefan Isler is Head of Software at Sony Advanced Visual Sensing, a Switzerland-based subsidiary of Sony Semiconductor Solutions Group, since 2020. At Sony AVS he coordinates the software developments, focusing on novel visual sensing solutions and systems. Previously, he was Head of Software at Insightness, a Zurich-based startup developing event-based sensors, systems and algorithms for a range of applications from drones to AR/VR which he joined after receiving his MSc in Mechanical Engineering from ETH Zurich.