Products-10K

Products-10K是由京东人工智能研究院创建的大型产品识别数据集，包含10,000个细粒度的SKU级别产品，这些产品是京东在线客户经常购买的。数据集涵盖了时尚、3C、食品、医疗保健、家居用品等多个类别，总计约150,000张图片。数据集中的图片分为店内照片和客户照片两类，其中店内照片背景清晰，客户照片背景复杂，有助于提高产品识别模型的泛化能力。数据集通过人工标注，确保标签的准确性，噪声率低于0.5%。该数据集主要用于计算机视觉社区中的日常产品细粒度SKU级别识别，旨在解决在线购物中产品识别的挑战。

Products-10K is a large-scale product recognition dataset created by the JD Artificial Intelligence Research Institute. It contains 10,000 fine-grained SKU-level products that are frequently purchased by JD's online customers. The dataset covers multiple categories including fashion, 3C, food, healthcare, household goods and others, with a total of approximately 150,000 images. The images in the dataset are divided into two categories: in-store photos and customer photos. In-store photos have clear backgrounds, while customer photos have complex backgrounds, which helps improve the generalization ability of product recognition models. The dataset is manually annotated to ensure label accuracy, with a noise rate of less than 0.5%. This dataset is primarily used for fine-grained SKU-level daily product recognition tasks in the computer vision community, aiming to address the challenges of product recognition in online shopping scenarios.