Event Vision-based Corner Detection with Count-normalized Multi-Layer Perceptron and Throughput Indicator

• DOI: 10.1016/j.compeleceng.2024.109432
• Publication Date: 2024-07-16T13:58:50Z
• Authors: Nan Zhang; Shupeng Han; Xiaoming Chen; Haodong Chen; Li Tan; Yuk Ying Chung
• Keywords: Leverage (statistics); Frame rate; Perceptron

The advancement of event cameras has sparked a revolution in imaging technology, presenting exciting opportunities for vision-based measurement tasks. Event operate on an innovative asynchronous principle, which offers several advantages over traditional cameras, including ultra-high dynamic range and exceptional temporal resolution. As result, excel challenging environments characterized by motion blur, overexposure, or underexposure, outperforming conventional frame-based cameras. However, the signal streams generated pose compatibility challenges with existing algorithms. This paper focuses specifically corner detection, critical task, tailored To address this challenge, we propose novel detectors that leverage advanced optimization techniques, enhanced time surface representations, multi-layer perceptron classifiers, throughput mechanism. Through rigorous experimentation, our method consistently shows lower projection errors compared to state-of-the-art methods across all datasets while also maintaining longer tracking times low-textured scenarios. Specifically, CMLP CMLP-T achieve average valid rate 83.38% 84.65%, respectively, DAVIS240C dataset collection, surpassing methods. We validate effectiveness proposed demonstrating their performance Furthermore, work contributes application machine learning processing tasks, providing insights into optimizing models unique characteristics