FoodLogoDet-1500
收藏资源简介:
FoodLogoDet-1500是由中国科学院计算技术研究所智能信息处理重点实验室创建的大型公开食品标志数据集,包含1,500个类别,约100,000张图像和145,400个手动标注的食品标志对象。该数据集通过广泛使用的购物应用Taobao和Jingdong以及Wikipedia构建食品标志类别列表,从多个搜索引擎收集食品标志图像,并通过人工检查确保图像包含相应的食品标志。FoodLogoDet-1500旨在支持食品标志检测研究,解决食品标志类别大规模和相似性问题,适用于健康饮食推荐、食品商标侵权检测等多种应用场景。
FoodLogoDet-1500 is a large-scale public food logo dataset created by the Key Laboratory of Intelligent Information Processing, Institute of Computing Technology, Chinese Academy of Sciences. It encompasses 1,500 categories, approximately 100,000 images, and 145,400 manually annotated food logo objects. The dataset constructs the food logo category list using the widely adopted shopping applications Taobao, Jingdong and Wikipedia, collects food logo images from multiple search engines, and verifies that the collected images contain the corresponding food logos via manual inspection. FoodLogoDet-1500 is intended to support food logo detection research, addressing the challenges of large-scale category counts and high inter-class similarity among food logo classes, and is applicable to multiple application scenarios such as healthy diet recommendation and food trademark infringement detection.

- 1FoodLogoDet-1500: A Dataset for Large-Scale Food Logo Detection via Multi-Scale Feature Decoupling Network中国科学院计算技术研究所智能信息处理重点实验室 · 2021年



