The MEVDT dataset was created to fill a critical gap in event-based computer vision research by supplying a high-quality, real-world labeled dataset. Intended to facilitate the development of advanced algorithms for object detection and tracking applications, MEVDT includes multi-modal traffic scene data with synchronized grayscale images and high-temporal-resolution event streams. Additionally, it provides annotations for object detection and tracking with class labels, pixel-precise bounding box coordinates, and unique object identifiers. The dataset is organized into directories containing sequences of images and event streams, comprehensive ground truth labels, fixed-duration event samples, and data indexing sets for training and testing. and To access and utilize the dataset, researchers need specific software or scripts compatible with the data formats included, such as PNG for grayscale images, CSV for event stream data, AEDAT for the encoded fixed-duration event samples, and TXT for annotations. Recommended tools include standard image processing libraries for PNG files and CSV or text parsers for event data. A specialized Python script for reading AEDAT files is available at: https://github.com/Zelshair/cstr-event-vision/blob/main/scripts/data_processing/read_aedat.py, which streamlines access to the encoded event sample data.
El Shair, Z. and Rawashdeh, S., 2024. MEVDT: Multi-Modal Event-Based Vehicle Detection and Tracking Dataset. Data In Brief (under review)., El Shair, Z. and Rawashdeh, S.A., 2022. High-temporal-resolution object detection and tracking using images and events. Journal of Imaging, 8(8), p.210., El Shair, Z. and Rawashdeh, S., 2023. High-temporal-resolution event-based vehicle detection and tracking. Optical Engineering, 62(3), pp.031209-031209., and El Shair, Z.A., 2024. Advancing Neuromorphic Event-Based Vision Methods for Robotic Perception Tasks (Doctoral dissertation, University of Michigan-Dearborn).
