Design and application of scientific dynamic vision sensor event camera

2 PhD or 1 Postdoc positions

(see this google doc link)

The Sensors Group at the Inst. of Neuroinformatics, UZH-ETH Zurich has two open PhD student (or one talented postdoc) positions for the the project SCIDVS funded by the Swiss National Science Foundation for 4 years.

SCIDVS will develop a high performance scientific dynamic vision sensor (DVS) event camera, and apply it to challenging problems in space survey, aerodynamics, and neuroimaging.

This optimized “scientific” event camera will enable scientific machine vision applications that are currently impossible with conventional frame-based camera technology. These applications are currently limited by the excessive data rate, limited dynamic range, and high latency of standard cameras. Neuromorphic “dynamic vision sensor” (DVS) event cameras can overcome these limitations by providing a sparse and quick, output consisting of only asynchronous brightness change events. Existing DVS cameras from academia and industry are designed as general-purpose devices with focus on small pixel size and low power consumption. Although feasibility studies have demonstrated their advantages for scientific vision applications, they are currently limited by noise and sensitivity. SCIDVS will design and fabricate an optimized DVS with lower noise, more sensitive event threshold, and higher quantum efficiency than existing DVS and fabricate it in an advanced CMOS image sensor process. There is expected to be an opportunity for interesting circuit design at all levels of the sensor, ranging from pixel to event readout.

To evaluate the SCIDVS camera, we will apply it (but first with existing event cameras) to three particularly challenging scientific vision applications, specifically space satellite tracking (SST), high-speed flow analysis by particle tracking velocimetry (PTV), and wide field neural activity imaging (NAI). SST demands low noise under low light conditions for nighttime observation and high event sensitivity under daytime conditions, while PTV pushes the boundaries for high speed, and NAI requires low noise and high sensitivity under low light conditions. A successful outcome will result in a scientific vision sensor with broad applicability for real-time analysis and continuous feedback control over imaging scales ranging from microscopic to astrometric.


    Position 1:
    Lead on DVS ASIC design. You will lead the design of the sensor in an advanced CMOS image sensor (CIS) process that provides high quantum efficiency and low noise buried photodiodes with microlenses or back side illumination.
    • Requirements: Mixed signal analog-digital CMOS design and testing experience, using Cadence/Mentor tools.
    • You will also participate in the application studies
  • Position 2: Lead on application studies in space satellite tracking, particle tracking velocimetry, and neural voltage imaging
    • Requirements: Experience with real time computer vision and background in mixed-signal or back-end IC design
    • You will also be helping with integration of the SCIDVS IC design, including layout and verification


To indicate interest in a position, send email to Prof. Tobi Delbruck ( with subject line containing “SCIDVS position”. You may wish to include the following material

  1. Your email should summarize briefly your background, accomplishment, and particular interest in SCIDVS
  2. Your CV including TOEFL, GRE, project/publication accomplishments, and 1 or 2 possible references.
  3. Your grade transcripts from undergraduate and masters programs
© 2020 Institut für Neuroinformatik