Asynchronous Event-Based Multikernel Algorithm for High-Speed Visual Features Tracking

TLDR

The paper introduces novel visual tracking methods that leverage the output of event‑based asynchronous neuromorphic dynamic vision sensors. The approach uses an asynchronous iterative framework that exploits spatial and temporal event correlations, applying multiple kernels (Gaussian, Gabor, etc.) and tracker pools to handle position, scale, and orientation changes while reducing per‑event complexity. Experiments show the method tracks multiple visual features in real time at several hundred kilohertz on a standard desktop PC, avoiding the N² per‑event cost of conventional kernel convolutions and demonstrating robust performance across kernel types.

Abstract

This paper presents a number of new methods for visual tracking using the output of an event-based asynchronous neuromorphic dynamic vision sensor. It allows the tracking of multiple visual features in real time, achieving an update rate of several hundred kilohertz on a standard desktop PC. The approach has been specially adapted to take advantage of the event-driven properties of these sensors by combining both spatial and temporal correlations of events in an asynchronous iterative framework. Various kernels, such as Gaussian, Gabor, combinations of Gabor functions, and arbitrary user-defined kernels, are used to track features from incoming events. The trackers described in this paper are capable of handling variations in position, scale, and orientation through the use of multiple pools of trackers. This approach avoids the N(2) operations per event associated with conventional kernel-based convolution operations with N × N kernels. The tracking performance was evaluated experimentally for each type of kernel in order to demonstrate the robustness of the proposed solution.