技術摘要（英）
Event cameras have a wider exposure range and a higher dynamic sensitivity than RGB cameras. However, we still need a method to correctly interpret the visual field for the following downstream applications. In this work, we develop an NN model for event-to-frame image reconstruction. While designed to be lightweight, this model has a formidable accuracy compared to state-of-the-art benchmarks. As a result, this model has better potential in the case of real-time AIoT applications.