List All Projects List Graduate Projects List Senior Design Projects List REU Projects Other Undergraduate Projects
Project # 53
Estimation of the frequency of periodic moving objects using high-speed event-based cameras
| Team Lead: | Aroudra Thakur |
| Other Team Members: | Christopher Tran | Tahsin Mostafiz | Diego Patino |
| Faculty Advisor(s): | Diego Patino |
| Department: | CSE |
| Project Type: | REU |
| Title: | Estimation of the frequency of periodic moving objects using high-speed event-based cameras |
| Abstract: | Event-based cameras are relatively new sensors to measure changes in brightness asynchronously, producing streams of events instead of standard images. These sensors offer high temporal resolution and low latency, making them well-suited for analyzing fast and periodic motion. In this work, we developed new computer vision algorithms for event-based data that are grounded in physical principles, specifically a formulation inspired by the heat equation. We model the spatio-temporal distribution of events using a heat-diffusion process, where event activity evolves similarly to the spreading of heat over a surface. We use this heat-inspired formulation to model the periodic motion of high-speed objects in visual scenes. By analyzing how the diffusion parameters change over time, we identify patterns that are strongly correlated with the frequency of periodic oscillations. Our results show that the parameters of our heat-based model encode information about motion frequency, allowing us to estimate the frequency directly from event streams. Overall, our work demonstrates that physics-inspired modeling can lead to robust and interpretable algorithms for periodic motion analysis in event-based vision systems. |
| Poster: | View Poster #53 (PDF) |
| Judges Scoring: | Submit Evaluation (judges only) |