{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/the-eurocity-persons-dataset-a-novel","title":"The EuroCity Persons Dataset: A Novel Benchmark for Object Detection","arxiv_id":"1805.07193","date":"2018-05-18","proceeding":null,"authors":["Markus Braun","Sebastian Krebs","Fabian Flohr","Dariu M. Gavrila"],"abstract":"Big data has had a great share in the success of deep learning in computer\nvision. Recent works suggest that there is significant further potential to\nincrease object detection performance by utilizing even bigger datasets. In\nthis paper, we introduce the EuroCity Persons dataset, which provides a large\nnumber of highly diverse, accurate and detailed annotations of pedestrians,\ncyclists and other riders in urban traffic scenes. The images for this dataset\nwere collected on-board a moving vehicle in 31 cities of 12 European countries.\nWith over 238200 person instances manually labeled in over 47300 images,\nEuroCity Persons is nearly one order of magnitude larger than person datasets\nused previously for benchmarking. The dataset furthermore contains a large\nnumber of person orientation annotations (over 211200). We optimize four\nstate-of-the-art deep learning approaches (Faster R-CNN, R-FCN, SSD and YOLOv3)\nto serve as baselines for the new object detection benchmark. In experiments\nwith previous datasets we analyze the generalization capabilities of these\ndetectors when trained with the new dataset. We furthermore study the effect of\nthe training set size, the dataset diversity (day- vs. night-time, geographical\nregion), the dataset detail (i.e. availability of object orientation\ninformation) and the annotation quality on the detector performance. Finally,\nwe analyze error sources and discuss the road ahead.","url_abs":"http://arxiv.org/abs/1805.07193v2","url_pdf":"http://arxiv.org/pdf/1805.07193v2.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[],"tasks":[{"task_slug":"benchmarking","task_name":"Benchmarking"},{"task_slug":"object","task_name":"Object"},{"task_slug":"object-detection","task_name":"Object Detection"},{"task_slug":"object-detection-1","task_name":"object-detection"}],"methods":[{"method_slug":"1x1-convolution","method_name":"1x1 Convolution"},{"method_slug":"convolution","method_name":"Convolution"},{"method_slug":"non-maximum-suppression","method_name":"Non Maximum Suppression"},{"method_slug":"position-sensitive-roi-pooling","method_name":"Position-Sensitive RoI Pooling"},{"method_slug":"r-fcn","method_name":"R-FCN"},{"method_slug":"ssd","method_name":"SSD"}],"datasets_introduced":[{"slug":"eurocity-persons","name":"EuroCity Persons","full_name":""}],"methods_introduced":[],"results":[],"syntology":{"syntology_url":null,"atlas_url":"https://app.syntology.ai/?focus=1805.07193","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}