Hairstyle, Facial Expression, and Clothing Dataset for Multi-Attribute Human Image Classification V1
Description
This dataset was developed for a modular deep learning pipeline for multi-attribute human image classification, focusing on three human appearance attributes: hairstyle, facial expression, and clothing style. It supports independent classification of each attribute while allowing their predictions to be combined within a unified human appearance profile. The dataset comprises 994 hairstyle images categorized as Curly, Straight, and Wavy; 2,579 facial-expression images categorized as Neutral and Smile; and 2,634 clothing-style images categorized as Casual and Formal. The hairstyle images were obtained from a publicly accessible Kaggle dataset. The facial-expression and clothing-style datasets were custom-developed for this study using images collected from Pexels, supplemented with photographs voluntarily provided by friends, family, and acquaintances with permission. These custom collections were developed to provide suitable data for the specific binary classification tasks and to introduce additional demographic and contextual variation. During dataset preparation, considerable effort was made in image selection, manual screening, class labeling, organization, and preprocessing. Images were selected to maintain clear class definitions and suitable visual quality while preserving natural variations in human appearance, including differences in gender, age, pose, background, and other visual characteristics. Each attribute-specific dataset was divided using a stratified 80:10:10 split into training, validation, and test sets. The released resource contains the processed images organized according to attribute, class, and dataset partition, corresponding to the data used in the reported experiments. The dataset is publicly shared to support research reproducibility, independent evaluation, and further development of lightweight and modular approaches for human attribute recognition and multi-attribute image classification.
Files
Institutions
- Comilla UniversityChittagong, Comilla