Facial Emotion Recognition Dataset with Three Expression Classes: Angry, Sad, Happy

This dataset was created by Group 10 – Bina Nusantara University @Bandung as part of the Artificial Intelligence course (AOL Project). It consists of a collection of human facial images categorized into three primary emotional classes: Angry, Sad, and Happy. All images were collected manually from Google Images and selected to represent real-world variations, including differences in lighting, camera angle, age, gender, and background. The dataset is designed to support research and development in Facial Emotion Recognition (FER), particularly for training deep learning models based on Convolutional Neural Networks (CNN) such as AlexNet. All images have been standardized through preprocessing steps, including resizing to 227×227 pixels, normalization, and labeling according to their respective emotional class. This dataset is intended for academic and research purposes, including machine learning experimentation, computer vision studies, and evaluation of emotion classification models based on facial imagery.