VIENA2
收藏资源简介:
VIENA2是由澳大利亚国立大学和Data61-CSIRO联合创建的大规模驾驶预测数据集,涵盖5种通用驾驶场景和25个不同的动作类别。该数据集包含超过15,000个全高清、5秒长的视频,总计超过225万帧,每帧都标注了动作标签。VIENA2通过使用GTA V视频游戏环境,模拟了多种驾驶条件、天气、时间和环境,以支持自动驾驶中的动作预测研究。数据集旨在通过提供丰富的多模态信息,解决自动驾驶中的复杂预测问题,如驾驶员行为、交通规则遵守、事故预测等。
VIENA2 is a large-scale driving prediction dataset jointly created by The Australian National University and Data61-CSIRO, covering 5 general driving scenarios and 25 distinct action categories. This dataset contains over 15,000 full high-definition 5-second-long videos, totaling more than 2.25 million frames, with each frame annotated with action labels. VIENA2 leverages the GTA V video game environment to simulate diverse driving conditions, weather patterns, times of day and environmental settings, supporting action prediction research in autonomous driving. This dataset aims to address complex prediction problems in autonomous driving, such as driver behavior, traffic rule compliance, accident prediction and more, by providing rich multimodal information.

- 1VIENA2: A Driving Anticipation Dataset澳大利亚国立大学 · 2018年



