Gender classification of PA-100K dataset, a Pedestrian Attribute dataset

This dataset is based on the work of Liu et al and their paper "Hydraplus-net: Attentive deep features for pedestrian analysis". In our work, we structure the images for a gender classification task based on the gender attribute annotated. Moreover, we pre-process the images to a 75x75 dimension that can be used by pre-trained deep learning models.

本数据集基于Liu等人的研究及其发表的论文《Hydraplus-net：面向行人分析的注意力深度特征》。本研究依据标注的性别属性，构建了用于性别分类任务的图像数据集。此外，我们将所有图像预处理至75×75的尺寸，以适配预训练深度学习模型的输入需求。