Searcharxiv⌕ Search

arXiv subjects

Zijie Ning

Publications and source records attributed to Zijie Ning.

3 recordsLinked to original sources

UnrealEvent: Synthetic Dataset Generation and Annotation for Event-Based Vision

Event-based cameras provide high temporal resolution, low power consumption, and a high dynamic range, but progress in event-based object detection remains constrained by the scarcity of annotated training data. To solve this issue, this paper introduces the tool UnrealEvent, a synthetic dataset-generation pipeline implemented in Unreal Engine 5.5. The pipeline combines a V2E-inspired event-frame generator with EasySynth semantic-rendering passes, allowing rendered level sequences to be converted into paired event representations and YOLO-format bounding-box annotations without manual labeling. Events are generated from per-pixel log-luminance differences between consecutive frames, with contrast thresholds sampled independently from a normal distribution to emulate the fabrication-induced threshold mismatch of real sensors. Using this pipeline, we provide a synthetic dataset and evaluate YOLOv8 detectors under synthetic, zero-shot real, and real fine-tuning training on eTraM. The results show strong in-domain synthetic performance but a substantial synthetic-to-real gap, especially for pedestrians. Fine-tuning on a small real subset improves real-domain performance, while the remaining gap highlights the importance of asset diversity and sensor calibration.

cs.CV↗

Event-Based Person Detection for Maritime UAV Search

Maritime search and rescue requires locating people in the water quickly, motivating the use of UAVs to survey the search area. During flight, camera motion and the longer exposures used in low light can blur small targets and hinder person detection. Event cameras record brightness changes as they occur, reducing the motion blur associated with longer exposure times. We investigate their use for maritime person detection by generating synthetic events using the v2e event simulator from public MOBDrone rescue videos and retaining the existing annotations. We compare the recurrent vision transformer model RVT-T with YOLOv8n and RT-DETR-R18 trained on exposure-matched images. On 843 test samples from held-out recordings, RVT-T achieves 15.63% person AP50:95, compared with 6.89% for YOLOv8n and 9.44% for RT-DETR-R18 under 100 ms synthetic exposure. Increasing exposure times from 50 to 200 ms progressively reduces frame-based detection accuracy, widening the gap to the event model. To evaluate the feasibility of on-board processing, we deploy our detector on a Jetson Orin Nano platform. The RVT-T model retains its accuracy and requires approximately 136 ms and 2.02 J per sample, allowing real-time processing on the UAV.

cs.CV↗

Raw2Event: Converting Raw Frame Camera into Event Camera

Event cameras offer unique advantages such as high temporal resolution, low latency, and high dynamic range, making them more and more popular for vision tasks under challenging light conditions. However, their high cost, limited resolution, and lack of features such as autofocus hinder their broad adoption, particularly for early-stage development and prototyping. In this work, we present Raw2Event, a complete hardware-software system that enables real-time event generation from low-cost raw frame-based cameras. By leveraging direct access to raw Bayer data and bypassing traditional image signal processors (ISP), our system is able to utilize the full potential of camera hardware, delivering higher dynamic range, higher resolution, and more faithful output than RGB-based frame-to-event converters. Built upon the DVS-Voltmeter model, Raw2Event features a configurable simulation framework optimized for deployment on embedded platforms. We further design a data acquisition pipeline that supports synchronized recording of raw, RGB, and event streams, facilitating downstream evaluation and dataset creation. Experimental results show that Raw2Event can generate event streams closely resembling those from real event cameras, while benefiting from higher resolution and autofocus capabilities. The system also supports user-intuitive parameter tuning, enabling flexible adaptation to various application requirements. Finally, we deploy the system on a Raspberry Pi for real-time operation, providing a scalable and cost-effective solution for event-based vision research and early-stage system development. The codes are available online: https://anonymous.4open.science/r/raw2event-BFF2/README.md.

cs.CV↗