头部姿态估计 | ViLab https://vilab.team/tag/%E5%A4%B4%E9%83%A8%E5%A7%BF%E6%80%81%E4%BC%B0%E8%AE%A1/ 头部姿态估计 Hugo Blox Builder (https://hugoblox.com)en-usSun, 29 Sep 2024 00:00:00 +0000 https://vilab.team/media/icon_hu2896232876136423579.png 头部姿态估计 https://vilab.team/tag/%E5%A4%B4%E9%83%A8%E5%A7%BF%E6%80%81%E4%BC%B0%E8%AE%A1/ Event-based head pose estimation: Benchmark and method https://vilab.team/publication/event-based-head-pose-estimation-benchmark-and-method/ Sun, 29 Sep 2024 00:00:00 +0000 https://vilab.team/publication/event-based-head-pose-estimation-benchmark-and-method/ <p>本文针对传统RGB方法在剧烈运动和极端光照下头部姿态估计困难的问题，引入事件相机的高时间分辨率与高动态范围优势。作者构建了两个大规模事件头部姿态数据集，包含282个序列，覆盖不同分辨率与场景；并提出事件头部姿态估计网络EV-HPE，设计了事件时空融合模块和事件运动感知注意力模块，有效结合事件流时空信息，提升姿态估计精度与鲁棒性。</p>