Gear8
收藏资源简介:
Gear8 是一个工业目标检测数据集,专门用于检测 Bosch 电动车窗电机减速器中 8 种不同类型的组件。该数据集在论文《MICA: Multi-Agent Industrial Coordination Assistant》和《Snap, Segment, Deploy: A Visual Data and Detection Pipeline for Wearable Industrial Assistants (SMC 2025)》中被提出和使用。数据集包含训练集和验证集,总大小约为 52GB,分为多个分片存档以便下载。训练集由 4 个分片组成,验证集和标注文件单独提供。数据集解压后的目录结构包括配置文件和分别存放训练与验证图像及标注的文件夹。该数据集适用于工业检测、齿轮检测等任务,特别适合使用 YOLO 等目标检测模型进行研究或应用开发。
Gear8 is an industrial object detection dataset dedicated to detecting 8 distinct types of components in Bosch power window motor reducers. This dataset was proposed and utilized in the papers *MICA: Multi-Agent Industrial Coordination Assistant* and *Snap, Segment, Deploy: A Visual Data and Detection Pipeline for Wearable Industrial Assistants (SMC 2025)*. It comprises a training set and a validation set, with a total size of approximately 52 GB, and is split into multiple chunked archives for download convenience. The training set consists of 4 chunks, while the validation set and annotation files are provided separately. The directory structure after decompression includes configuration files and separate folders for storing training and validation images as well as their annotations. This dataset is suitable for tasks such as industrial detection and gear detection, and is particularly well-suited for research or application development using object detection models like YOLO.




