eProduct

eProduct是由eBay Inc.创建的大规模产品图像数据集，旨在解决电子商务领域中的产品识别挑战。该数据集包含250万张产品图片，涵盖16个元类别和1000个叶类别，用于支持自监督学习、弱监督学习和多模态学习等领域的研究。数据集的创建过程涉及从eBay3500内部数据集中随机选择和剪裁类别，以确保数据分布的平衡。eProduct不仅用于模型训练，还包含10,000个查询图像和110万张索引图像，用于视觉搜索评估，旨在提高产品识别的准确性和效率。eProduct is a large-scale product image dataset created by eBay Inc., which aims to address product recognition challenges in the e-commerce domain. This dataset contains 2.5 million product images, covering 16 meta-categories and 1,000 leaf categories, and supports research in fields such as self-supervised learning, weakly-supervised learning, and multimodal learning. The dataset creation process involves randomly selecting and pruning categories from eBay3500's internal dataset to ensure a balanced data distribution. In addition to being used for model training, eProduct also includes 10,000 query images and 1.1 million index images for visual search evaluation, with the goal of improving the accuracy and efficiency of product recognition.

搜集汇总

背景与挑战

背景概述

eProduct是eBay Inc.创建的大规模产品图像数据集，包含250万张图片，覆盖16个元类别和1000个叶类别，旨在解决电子商务产品识别问题。该数据集支持自监督、弱监督和多模态学习研究，并包含查询和索引图像用于视觉搜索评估，以提高产品识别的准确性和效率。