MUTFER2024

MUTFER2024 was developed at Mangosuthu University of Technology to provide facial emotion data from South African participants under real-world conditions. The dataset includes 13,032 images across seven emotion categories, captured without constraints on lighting, backgrounds, or accessories. Such variability aligns with the need for robustness in real-world FER systems, as highlighted in studies on mobile and uncontrolled environments.

MUTFER2024数据集由曼戈苏图理工大学（Mangosuthu University of Technology）研发，旨在采集南非参与者在真实场景下的面部表情数据。该数据集包含13032张图像，涵盖7种表情类别，拍摄过程未对光照条件、背景环境或佩戴配饰施加任何限制。此类数据的变异性恰好契合真实场景下面部表情识别（Facial Expression Recognition，FER）系统对鲁棒性的需求，这一要点在面向移动设备与非受控环境的相关研究中已得到明确阐释。

搜集汇总

背景与挑战

背景概述

MUTFER2024是由Mangosuthu University of Technology开发的面部情绪识别数据集，包含13,032张图像，覆盖七种情绪类别，所有图像在真实世界条件下采集，无光照、背景或配饰限制，旨在提高FER系统在非受控环境中的鲁棒性。