SFEW (Static Facial Expression in the Wild)

野生静态面部表情 (SFEW) 数据集是用于面部表情识别的数据集。它是通过基于面部点聚类计算关键帧从 AFEW 数据库中选择静态帧而创建的。最常用的版本 SFEW 2.0 是 EmotiW 2015 中 SReco 子挑战的基准数据。SFEW 2.0 分为三组：Train（958 个样本）、Val（436 个样本）和 Test（372 个样本）。每个图像都被分配到七种表情类别之一，即愤怒、厌恶、恐惧、中性、快乐、悲伤和惊讶。训练集和验证集的表达标签是公开的，而测试集的表达标签则由挑战组织者保留。

The Static Facial Expressions in the Wild (SFEW) dataset is a benchmark dataset for facial expression recognition tasks. It was developed by selecting static frames from the AFEW database through key frame computation based on facial point clustering. The most widely used version, SFEW 2.0, served as the benchmark data for the SReco sub-challenge at EmotiW 2015. SFEW 2.0 is divided into three subsets: Train (958 samples), Val (436 samples), and Test (372 samples). Each image is assigned to one of seven facial expression categories: anger, disgust, fear, neutral, happiness, sadness, and surprise. Expression labels for the training and validation sets are publicly available, while those for the test set are withheld by the challenge organizers.