cs.CVMar 23, 2026

Benchmarking Recurrent Event-Based Object Detection for Industrial Multi-Class Recognition on MTevent

Authors: Lokeshwaran Manohar, Moritz Roidl

Organizations: Chair of Material Handling and Warehousing, TU Dortmund University, Dortmund, Germany

Read PDF
Ask AI
Cite

Copy citation

DOINot available for this paper

Abstract

Event cameras are attractive for industrial robotics because they provide high temporal resolution, high dynamic range, and reduced motion blur. However, most event-based object detection studies focus on outdoor driving scenarios or limited class settings. In this work, we benchmark recurrent ReYOLOv8s on MTevent for industrial multi-class recognition and use a non-recurrent YOLOv8s variant as a baseline to analyze the effect of temporal memory. On the MTevent validation split, the best scratch recurrent model (C21) reaches 0.285 mAP50, corresponding to a 9.6% relative improvement over the non-recurrent YOLOv8s baseline (0.260). Event-domain pretraining has a stronger effect: GEN1-initialized fine-tuning yields the best overall result of 0.329 mAP50 at clip length 21, and unlike scratch training, GEN1-pretrained models improve consistently with clip length. PEDRo initialization drops to 0.251, indicating that mismatched source-domain pretraining can be less effective than training from scratch. Persistent failure modes are dominated by class imbalance and human-object interaction. Overall, we position this work as a focused benchmarking and analysis study of recurrent event-based detection in industrial environments.

Figures & tables

ModelRecurrentInitClipmAP50mAP50-95
YOLOv8s baselineNoScratch10.2600.138
ReYOLOv8sYesScratch30.2620.148
ReYOLOv8sYesScratch70.2710.152
ReYOLOv8sYesScratch110.2610.141
ReYOLOv8sYesScratch210.2850.155
ReYOLOv8s (GEN1 init)YesGEN130.2930.157
TABLE I: ABLATION RESULTS ON THE MTEVENT VALIDATION SPLIT COMPARING NON-RECURRENT AND RECURRENT MODELS, TEMPORAL CLIP LENGTH, AND EVENT-DOMAIN PRETRAINING.
Fig. 1: Validation performance on MTevent across non-recurrent and recurrent settings, temporal clip lengths, and pretraining configurations. YOLO denotes the non-recurrent YOLOv8s baseline, ReY-C nn denotes ReYOLOv8s trained from scratch with clip length nn , and GEN1/PEDRo-C nn denote ReYOLOv8s fine-tuned from pretrained weights at clip length nn . GEN1 initialization yields the best overall result at C21. Numerical values are reported in Table I .
Fig. 2: Qualitative zero-shot transfer examples on MTevent. The GEN1-pretrained model does not transfer reliably to MTevent humans, whereas the PEDRo-pretrained model produces a stronger human detection. The fine-tuned MTevent model detects the human with the correct target label.
More in this topic

Event-Based Object Detection

Event-based object detection localizes and classifies objects from asynchronous camera events, aiming for fast, robust perception with low latency and power.

See all 36 papers

Explore similar work

CardsList

  1. Efficient Multi-Timescale Event Representations for Feed-Forward Object Detection

    Sep 4, 2026Fredrik Lundell, Per-Erik Forssen, Mårten Wadenbäck +1Event-Based VisionEvent-Based Object Detection

    Department of Electrical Engineering, Linköping University, Sweden · Saab AB, Linköping, Sweden

  2. Sensor Generalization for Adaptive Sensing in Event-based Object Detection via Joint Distribution Training

    Feb 26, 2026Aheli Saha, René Schuster, Didier StrickerAdaptive SensingEvent-Based Vision

    Deutsches Forschungszentrum für Künstliche Intelligenz, Kaiserslautern, Germany

  3. A Multimodal RGB and Events Dataset for Hand Detection in First-Person View

    Jun 9, 2026Bharghav Kota, Yulia SandamirskayaEvent-Based VisionEvent-Based Object Detection

    Zurich University of Applied Sciences, Wädenswil, Switzerland · Zurich University of Applied Sciences Wädenswil, Switzerland