Event Sensing
We investigate efficient processing of event-based sensor data for low-power, real-time perception. Event cameras report changes asynchronously, allowing processing to focus on activity in the scene.
Research directions
Our research connects sensor data processing with spiking neural networks (SNNs) and embedded hardware. Topics include:
- Event filtering and compression to reduce data movement and processing cost.
- Real-time object and motion detection from event streams.
- SNN processing and FPGA acceleration for low-power perception.
- Embedded sensing systems that combine event and thermal cameras for human detection.
Students can contribute through Python software, FPGA and VLSI design, or combined hardware and software implementation. See the event sensor processing thesis theme.
Prototypes and demonstrations
Our 2026 Open Lab materials introduce real-time moving-object detection using an FPGA and event camera and human detection using a Raspberry Pi and thermal camera.
- Summer Open Lab: sensor setups and demonstration video.
- Fall Open Lab: low-power sensing demonstrations.
- Event Processing Course: related teaching and practical work.
Related research
- IRIS project: related work in event sensing.
- Low-Power Neuromorphic Computing: approximate computing and stacked memory for low-power SNNs.
- HYMEX: hybrid memory architectures for energy-efficient AI.
- Publications: the group's research papers, including work on thermal and event-camera sensing.
Work with us
Explore graduation thesis themes or contact us about event sensing and embedded processing.