DriveGaze
收藏资源简介:
DriveGaze数据集是由大连理工大学创建的,旨在用于驾驶员状态识别研究。该数据集包含从传统摄像头生成的强度帧和相应的合成动态视觉传感器(DVS)事件,涵盖了不同年龄、种族和性别的数据。数据集的创建过程包括从视频帧生成合成DVS事件,并使用注意力驱动状态网络(ADSN)进行处理。DriveGaze数据集主要应用于驾驶员状态识别领域,旨在通过低成本的传统摄像头捕捉的眼睛特征来识别驾驶员的状态,解决传统摄像头在有限光照条件下无法有效解析时间信息的问题。
The DriveGaze dataset was developed by Dalian University of Technology for driver state recognition research. It includes intensity frames generated by traditional cameras and corresponding synthetic dynamic vision sensor (DVS) events, covering data from diverse age groups, ethnicities and genders. The construction of the dataset involves generating synthetic DVS events from video frames and processing them using the Attention-Driven State Network (ADSN). Primarily applied in the field of driver state recognition, the DriveGaze dataset aims to identify driver states via eye features captured by low-cost traditional cameras, solving the problem that traditional cameras fail to effectively parse temporal information under limited lighting conditions.

- 1DriveGazen: Event-Based Driving Status Recognition using Conventional Camera大连理工大学 · 2024年



