远程学习情感与生理数据集(RLAP)
收藏资源简介:
远程学习情感与生理数据集(RLAP)是由华中师范大学人工智能教育学部创建的大型同步无损格式数据集。该数据集包含来自58名受试者的超过32小时(353万帧)视频,使用PhysRecorder和低成本网络摄像头收集。RLAP数据集旨在解决远程生理信号恢复中的挑战,特别是在视频压缩和信号同步方面。通过在RLAP上训练,包括DeepPhys、TS-CAN、PhysNet和PhysFormer在内的多种神经算法实现了更高的泛化能力。该数据集适用于远程学习和医疗领域的情感计算和生理信号分析,有助于开发更高效的算法和模型。
The Remote Learning Affective and Physiological Dataset (RLAP) is a large-scale synchronized lossless-format dataset developed by the Faculty of Artificial Intelligence and Education, Central China Normal University. This dataset contains over 32 hours (3.53 million frames) of video from 58 subjects, collected using PhysRecorder and low-cost webcams. The RLAP dataset aims to address the challenges in remote physiological signal recovery, particularly regarding video compression and signal synchronization. Multiple neural algorithms including DeepPhys, TS-CAN, PhysNet, and PhysFormer have demonstrated enhanced generalization capabilities after being trained on the RLAP dataset. This dataset is suitable for affective computing and physiological signal analysis in the fields of remote learning and healthcare, and supports the development of more efficient algorithms and models.

- 1PhysBench: A Benchmark Framework for rPPG with a New Dataset and Baseline华中师范大学人工智能教育学部 · 2023年



