THEODORE

THEODORE数据集是由开姆尼茨工业大学电气工程与信息技术学院创建的大型室内数据集，包含100,000张高分辨率的多样化鱼眼图像，涵盖16个类别。数据集通过Unity3D游戏引擎生成3D虚拟环境，包括生活房间、不同的人类角色和室内纹理，并提供语义分割、实例掩码和边界框注释。该数据集适用于深度转移学习，特别是在全向图像的对象检测和语义分割领域，旨在提高现有CNN模型在全向图像上的准确性，应用于自主系统的视觉里程计导航、公共交通服务中的个人安全和虚拟现实等领域。

The THEODORE dataset is a large-scale indoor dataset developed by the Faculty of Electrical Engineering and Information Technology, Chemnitz University of Technology. It contains 100,000 high-resolution and diverse fisheye images spanning 16 categories. The dataset generates 3D virtual environments through the Unity3D game engine, which includes living rooms, various human characters and indoor textures, and provides annotations for semantic segmentation, instance masks and bounding boxes. This dataset is suitable for deep transfer learning, especially in the fields of object detection and semantic segmentation on omnidirectional images. It aims to improve the accuracy of existing CNN models when processing omnidirectional images, with applications in visual odometry navigation for autonomous systems, personal safety in public transportation services, virtual reality and other related domains.