VWise

VWise数据集是由联邦计算机中心，伯南布哥联邦大学等机构共同创建，专注于拉丁美洲地区的道路类型分类。该数据集包含521个视频片段，覆盖城市、工业和农村环境，涉及六种道路类型。创建过程中，数据来源于YouTube和视频库，通过特定的查询和筛选获取。VWise旨在解决现有数据集的地理偏差问题，特别是在自动驾驶和驾驶辅助系统中的应用，以提升模型在拉丁美洲地区的适应性和准确性。

The VWise dataset was co-developed by institutions including the Federal Computer Center and Federal University of Pernambuco, focusing on road type classification in the Latin American region. This dataset comprises 521 video clips covering urban, industrial and rural environments, and encompasses six road types. During its creation, data was collected from YouTube and video libraries via specific queries and filtering processes. The VWise dataset is designed to address the geographic bias issue prevalent in existing datasets, particularly for applications in autonomous driving and driving assistance systems, thereby enhancing the adaptability and accuracy of models when deployed in the Latin American region.