FEAFA+

FEAFA由两个子数据集组成: FEAFA (子数据集A) 和重新标记的DISFA (子数据集B)。
FEAFA: 有127个带有99,356帧标记的排列序列。在现实世界中记录了122名参与者，包括儿童，年轻人和老年人。要求所有参与者在面部表情数据记录和标记协议中进行所需的面部动作。
重新标记的DISFA: 有27个自发序列，其中130,828个帧重新标记。原始数据是在特定环境中使用特定的立体视觉系统记录的，参与者主要是各种种族的年轻人。
此外，使用我们开发的表达定量工具手动标记了总共230,184帧，以量化9个对称FACS动作单元、10个不对称 (单边) FACS动作单元、2个对称FACS动作描述符和2个不对称FACS动作描述符，并且每个动作单元或动作描述符都带有0到1之间的浮点数。

FEAFA consists of two subdatasets: FEAFA (Subdataset A) and the re-labeled DISFA (Subdataset B).
FEAFA contains 127 posed facial expression sequences with 99,356 annotated frames. It was recorded from 122 participants across real-world scenarios, including children, young adults and the elderly. All participants were required to perform the required facial movements in accordance with the facial expression data recording and annotation protocol.
The re-labeled DISFA includes 27 spontaneous facial expression sequences, with a total of 130,828 frames being re-annotated. The original data was recorded using a specific stereo vision system in a specific environment, and the participants were mainly young adults of various ethnicities.
In addition, a total of 230,184 frames were manually annotated using our developed expression quantification tool to quantify 9 symmetric FACS Action Units (AUs), 10 asymmetric (unilateral) FACS Action Units, 2 symmetric FACS Action Descriptors and 2 asymmetric FACS Action Descriptors, with each action unit or descriptor assigned a floating-point value ranging from 0 to 1.