MUTFER2024: A new dataset for South African emotion recognition

IF 1 Q3 MULTIDISCIPLINARY SCIENCES

Data in Brief Pub Date : 2025-04-28 DOI:10.1016/j.dib.2025.111592

Rogerant Tshibangu , Jules R Tapamo

{"title":"MUTFER2024: A new dataset for South African emotion recognition","authors":"Rogerant Tshibangu , Jules R Tapamo","doi":"10.1016/j.dib.2025.111592","DOIUrl":null,"url":null,"abstract":"<div><div>Facial Emotion Recognition (FER) plays a critical role in applications such as human-computer interaction, security, and healthcare. The effectiveness of FER systems largely depends on the quality and diversity of the datasets used for training and evaluation. However, existing FER datasets often lack adequate representation of African populations, leading to racial biases in recognizing emotions across diverse ethnic groups. This issue arises from the predominance of Western-centric datasets used in training FER systems, which results in inaccurate and biased outcomes when applied to African or non-Caucasian faces.</div><div>To address this limitation, we introduce MUTFER2024, a novel dataset developed at Mangosuthu University of Technology. MUTFER2024 aims to minimize racial bias in FER systems by providing an extensive collection of facial emotion images from African participants. The dataset comprises 13,032 images collected from 300 individuals, including students and staff members, and is categorized into seven emotion classes: happy, sad, angry, surprised, neutral, disgusted, and fearful.</div><div>This paper details the methodology employed in data collection, segmentation, and categorization. Facial emotion images were gathered through structured submission protocols to ensure diversity in expressions. Subsequently, the images were meticulously segmented and categorized into the specified emotion classes. Data were collected under real-world conditions using mobile and computer cameras. The dataset is hosted on GitHub and can be used to train emotion recognition models for underrepresented African populations.</div></div>","PeriodicalId":10973,"journal":{"name":"Data in Brief","volume":"60 ","pages":"Article 111592"},"PeriodicalIF":1.0000,"publicationDate":"2025-04-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Data in Brief","FirstCategoryId":"1085","ListUrlMain":"https://www.sciencedirect.com/science/article/pii/S2352340925003245","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q3","JCRName":"MULTIDISCIPLINARY SCIENCES","Score":null,"Total":0}

引用次数: 0

Abstract

Facial Emotion Recognition (FER) plays a critical role in applications such as human-computer interaction, security, and healthcare. The effectiveness of FER systems largely depends on the quality and diversity of the datasets used for training and evaluation. However, existing FER datasets often lack adequate representation of African populations, leading to racial biases in recognizing emotions across diverse ethnic groups. This issue arises from the predominance of Western-centric datasets used in training FER systems, which results in inaccurate and biased outcomes when applied to African or non-Caucasian faces.

To address this limitation, we introduce MUTFER2024, a novel dataset developed at Mangosuthu University of Technology. MUTFER2024 aims to minimize racial bias in FER systems by providing an extensive collection of facial emotion images from African participants. The dataset comprises 13,032 images collected from 300 individuals, including students and staff members, and is categorized into seven emotion classes: happy, sad, angry, surprised, neutral, disgusted, and fearful.

This paper details the methodology employed in data collection, segmentation, and categorization. Facial emotion images were gathered through structured submission protocols to ensure diversity in expressions. Subsequently, the images were meticulously segmented and categorized into the specified emotion classes. Data were collected under real-world conditions using mobile and computer cameras. The dataset is hosted on GitHub and can be used to train emotion recognition models for underrepresented African populations.

查看原文本刊更多论文

MUTFER2024：南非情绪识别的新数据集

面部情绪识别（FER）在人机交互、安全和医疗保健等应用中起着至关重要的作用。FER系统的有效性在很大程度上取决于用于训练和评估的数据集的质量和多样性。然而，现有的FER数据集往往缺乏对非洲人口的充分代表，导致在识别不同种族群体的情绪时存在种族偏见。这一问题源于在训练FER系统中使用的以西方为中心的数据集占主导地位，这导致在应用于非洲或非白种人面孔时结果不准确和有偏见。为了解决这一限制，我们引入了MUTFER2024，这是一个由Mangosuthu理工大学开发的新数据集。MUTFER2024旨在通过提供大量来自非洲参与者的面部表情图像，最大限度地减少fers系统中的种族偏见。该数据集包括从300个人（包括学生和教职员工）收集的13032张图像，并将其分为七种情绪类别：快乐、悲伤、愤怒、惊讶、中性、厌恶和恐惧。本文详细介绍了在数据收集、分割和分类中采用的方法。通过结构化提交协议收集面部情绪图像，以确保表情的多样性。随后，图像被精心分割并分类到指定的情感类别中。数据是在真实环境下使用手机和电脑相机收集的。该数据集托管在GitHub上，可用于为代表性不足的非洲人口训练情感识别模型。

本文章由计算机程序翻译，如有差异，请以英文原文为准。

求助全文

约1分钟内获得全文求助全文

来源期刊

Data in Brief MULTIDISCIPLINARY SCIENCES-

CiteScore

3.10

自引率

0.00%

发文量

996

审稿时长

70 days

期刊介绍： Data in Brief provides a way for researchers to easily share and reuse each other''s datasets by publishing data articles that: -Thoroughly describe your data, facilitating reproducibility. -Make your data, which is often buried in supplementary material, easier to find. -Increase traffic towards associated research articles and data, leading to more citations. -Open up doors for new collaborations. Because you never know what data will be useful to someone else, Data in Brief welcomes submissions that describe data from all research areas.