The EuroCity Persons Dataset: A Novel Benchmark for Object Detection

Abstract

Big data has had a great share in the success of deep learning in computer vision. Recent works suggest that there is significant further potential to increase object detection performance by utilizing even bigger datasets. In this paper, we introduce the EuroCity Persons dataset, which provides a large number of highly diverse, accurate and detailed annotations of pedestrians, cyclists and other riders in urban traffic scenes. The images for this dataset were collected on-board a moving vehicle in 31 cities of 12 European countries. With over 238200 person instances manually labeled in over 47300 images, EuroCity Persons is nearly one order of magnitude larger than person datasets used previously for benchmarking. The dataset furthermore contains a large number of person orientation annotations (over 211200). We optimize four state-of-the-art deep learning approaches (Faster R-CNN, R-FCN, SSD and YOLOv3) to serve as baselines for the new object detection benchmark. In experiments with previous datasets we analyze the generalization capabilities of these detectors when trained with the new dataset. We furthermore study the effect of the training set size, the dataset diversity (day- vs. night-time, geographical region), the dataset detail (i.e. availability of object orientation information) and the annotation quality on the detector performance. Finally, we analyze error sources and discuss the road ahead.

arXiv commentSubmitted to IEEE Trans. on Pattern Analysis and Machine Intelligence

References in corpus (8)

• Very Deep Convolutional Networks for Large-Scale Image Recognition
• Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks
• YOLOv3: An Incremental Improvement
• Caffe: Convolutional Architecture for Fast Feature Embedding
• R-FCN: Object Detection via Region-based Fully Convolutional Networks
• What makes for effective detection proposals?
• What makes ImageNet good for transfer learning?
• Scale-aware Fast R-CNN for Pedestrian Detection

Cited by in corpus (36)

• Deep Multi-modal Object Detection and Semantic Segmentation for Autonomous Driving: Datasets, Methods, and Challenges
• A Survey of Deep Learning-based Object Detection
• Towards Large-Scale Small Object Detection: Survey and Benchmarks
• Perception and Sensing for Autonomous Vehicles Under Adverse Weather Conditions: A Survey
• Object Detection in 20 Years: A Survey
• A Review and Comparative Study on Probabilistic Object Detection in Autonomous Driving
• Anomaly Detection in Autonomous Driving: A Survey
• From Handcrafted to Deep Features for Pedestrian Detection: A Survey
• CNN based Road User Detection using the 3D Radar Cube
• Recent Advances in Object Detection in the Age of Deep Convolutional Neural Networks
• A Survey on Approximate Edge AI for Energy Efficient Autonomous Driving Services
• A Novel CNN-based Method for Accurate Ship Detection in HR Optical Remote Sensing Images via Rotated Bounding Box
• TJU-DHD: A Diverse High-Resolution Dataset for Object Detection
• Uncertainty Estimation in One-Stage Object Detection
• An Application-Driven Conceptualization of Corner Cases for Perception in Highly Automated Driving
• Co-Design of Embodied Intelligence: A Structured Approach
• Driving Datasets Literature Review
• Adapted Center and Scale Prediction: More Stable and More Accurate
• E-Scooter Rider Detection and Classification in Dense Urban Environments
• Vulnerable Road User Detection and Safety Enhancement: A Comprehensive Survey
• PedHunter: Occlusion Robust Pedestrian Detector in Crowded Scenes
• RGB-D Railway Platform Monitoring and Scene Understanding for Enhanced Passenger Safety
• Generalizable Pedestrian Detection: The Elephant In The Room
• VEGA: Towards an End-to-End Configurable AutoML Pipeline
• Robustness in Compressed Neural Networks for Object Detection
• STRIDE: Street View-based Environmental Feature Detection and Pedestrian Collision Prediction
• The Impact of Partial Occlusion on Pedestrian Detectability
• When We First Met: Visual-Inertial Person Localization for Co-Robot Rendezvous
• WiderPerson: A Diverse Dataset for Dense Pedestrian Detection in the Wild
• Vehicle Telematics Via Exteroceptive Sensors: A Survey
• Detecting Reflections by Combining Semantic and Instance Segmentation
• SHD360: A Benchmark Dataset for Salient Human Detection in 360° Videos
• Weakly Supervised Dataset Collection for Robust Person Detection
• AFAN: Augmented Feature Alignment Network for Cross-Domain Object Detection
• PANDA: A Gigapixel-level Human-centric Video Dataset
• A lightweight target detection algorithm based on Mobilenet Convolution