FER2025: A Deep Learning Approach to Facial Emotion Recognition with Gender Classification Using CNN

This dataset was developed for a deep learning project focused on facial emotion recognition with gender classification using Convolutional Neural Networks (CNNs). It comprises a total of 7,386 images and 11,253 annotations, collected exclusively from reputable sources offering copyright-free or royalty-free images, such as Unsplash, iStock, Shutterstock, Getty Images etc. These platforms explicitly permit usage for personal and commercial purposes without requiring attribution.

The dataset includes 12 distinct classes, corresponding to six emotion categories (e.g., happy, sad, angry, fear, surprised, sleepy), each labeled separately for male and female subjects, resulting in a total of 12 unique class labels.

To facilitate effective model training and evaluation, the dataset is organized into three subsets:
Training Set: 70% of the data
Validation Set: 20% of the data
Test Set: 10% of the data

The metadata_FER2025.csv file lists each image of a facial emotion with its name, set type (Train/Valid/Test), Class ID, Class Name, and bounding box details (X, Y, Width, Height). It organizes the dataset for efficient object detection model training.

All annotations were carefully curated to ensure compliance with licensing terms and legal usage rights.