| { |
| "@context": { |
| "@language": "en", |
| "@vocab": "https://schema.org/", |
| "citeAs": "cr:citeAs", |
| "column": "cr:column", |
| "conformsTo": "dct:conformsTo", |
| "cr": "http://mlcommons.org/croissant/", |
| "rai": "http://mlcommons.org/croissant/RAI/", |
| "data": { |
| "@id": "cr:data", |
| "@type": "@json" |
| }, |
| "dataType": { |
| "@id": "cr:dataType", |
| "@type": "@vocab" |
| }, |
| "dct": "http://purl.org/dc/terms/", |
| "examples": { |
| "@id": "cr:examples", |
| "@type": "@json" |
| }, |
| "extract": "cr:extract", |
| "field": "cr:field", |
| "fileProperty": "cr:fileProperty", |
| "fileObject": "cr:fileObject", |
| "fileSet": "cr:fileSet", |
| "format": "cr:format", |
| "includes": "cr:includes", |
| "isLiveDataset": "cr:isLiveDataset", |
| "jsonPath": "cr:jsonPath", |
| "key": "cr:key", |
| "md5": "cr:md5", |
| "parentField": "cr:parentField", |
| "path": "cr:path", |
| "recordSet": "cr:recordSet", |
| "references": "cr:references", |
| "regex": "cr:regex", |
| "repeated": "cr:repeated", |
| "replace": "cr:replace", |
| "sc": "https://schema.org/", |
| "separator": "cr:separator", |
| "source": "cr:source", |
| "subField": "cr:subField", |
| "transform": "cr:transform", |
| "wd": "https://www.wikidata.org/wiki/" |
| }, |
| "@type": "sc:Dataset", |
| "conformsTo": "http://mlcommons.org/croissant/1.0", |
| "name": "Refined Annotations with Soft Labels for Common Object Detection Benchmarks", |
| "url": "https://huggingface.co/datasets/JKlees/Uncertainty_Aware_Object_Detection", |
| "creator": [ |
| { |
| "@type": "Person", |
| "givenName": "Anonymous", |
| "familyName": "Anonymous", |
| "name": "Anonymous, Anonymous" |
| } |
| ], |
| "description": "These datasets contain refined annotations for three popular object detection benchmark datasets: COCO 2017 validation data, Pascal VOC 2012 detection data for both training and validation split, and KITTI 2D object detection training data. Each annotation consists of a bounding box (xmin, ymin, xmax, ymax) and a soft label annotation ('postprocessed_soft_label') as well as original annotator responses for object categories by annotator ID, including answer times.", |
| "keywords": [ |
| "Computer and Information Science", |
| "Object Detection" |
| ], |
| "license": "http://creativecommons.org/licenses/by-nc-sa/4.0", |
| "datePublished": "2026-10-19", |
| "dateModified": "2026-10-19", |
| "includedInDataCatalog": { |
| "@type": "DataCatalog", |
| "name": "Hugging Face Datasets", |
| "url": "https://huggingface.co" |
| }, |
| "publisher": { |
| "@type": "Organization", |
| "name": "Hugging Face" |
| }, |
| "version": "1.0", |
| "citeAs": "@inproceedings{penquitt2026object, author = {Penquitt, Sarina and Klees, Jonathan and van Betteray, Antonia and Jashnieh, Parssa and Stehr, Peter and Rottmann, Matthias and Schmarje, Lars} ,title = {{Object Detection Benchmarks are Incomplete: The Role of Label Errors and Annotation Uncertainty}}, booktitle = {Advances in Neural Information Processing Systems Evaluations and Datasets Track}, year = {2026}, }", |
| "rai:dataLimitations": "These datasets include more objects than the original annotations, as well as soft labels based on at least 11 annotator responses. However, there is no perfect data and objects are also missing in our annotations, we elaborate on this in Section 4 of the paper.", |
| "rai:dataBiases": "We noticed a confirmation bias of annotators when answering the bounding box verification task, answers tend to accept the proposed class. We aim at correcting this with existing methods but human responses are subject to annotator reliability.", |
| "rai:personalSensitiveInformation": "No personal or sensitive information is included in the dataset as we only publish the annotations, not the images.", |
| "rai:dataUseCases": "The dataset can be used for training and evaluating object detection models, in particular with respect to uncertainty quantification, as well as for research on annotation quality, annotator reliability, and human-in-the-loop annotation methods.", |
| "rai:dataSocialImpact": "There is no direct societal impact of our refined annotations. The datasets can be used to improve object detection models, which have a wide range of applications in areas such as autonomous driving, surveillance, and robotics. It is important to consider the potential ethical implications of these applications and to ensure that the dataset is used responsibly.", |
| "rai:hasSyntheticData": "No synthetic data is included in the dataset, all annotations are based on human responses.", |
| "distribution": [ |
| { |
| "@type": "cr:FileObject", |
| "@id": "soft_COCO_2017_val.json", |
| "name": "soft_COCO_2017_val.json", |
| "encodingFormat": "application/json", |
| "md5": "ea70d5d0ac71e585f9279c0ebf0340e1", |
| "contentSize": "962000000", |
| "description": "Annotations for COCO 2017 validation data (5,000 images)", |
| "contentUrl": "https://huggingface.co/datasets/JKlees/Uncertainty_Aware_Object_Detection/blob/main/soft_COCO_2017_val.json" |
| }, |
| { |
| "@type": "cr:FileObject", |
| "@id": "soft_Kitti_2D_train.json", |
| "name": "soft_Kitti_2D_train.json", |
| "encodingFormat": "application/json", |
| "md5": "90ebbdf43097954204f624e0cfd4758f", |
| "contentSize": "393000000", |
| "description": "Annotations for the KITTI 2D Object Detection task (7,481 images)", |
| "contentUrl": "https://huggingface.co/datasets/JKlees/Uncertainty_Aware_Object_Detection/blob/main/soft_Kitti_2D_train.json" |
| }, |
| { |
| "@type": "cr:FileObject", |
| "@id": "soft_PascalVOC_2012_detection_train.json", |
| "name": "soft_PascalVOC_2012_detection_train.json", |
| "encodingFormat": "application/json", |
| "md5": "dc38a99a3012ff24a05c4a869f7c7114", |
| "contentSize": "128000000", |
| "description": "Annotations for the Pascal VOC 2012 detection task, training data (5,717 images)", |
| "contentUrl": "https://huggingface.co/datasets/JKlees/Uncertainty_Aware_Object_Detection/blob/main/soft_PascalVOC_2012_detection_train.json" |
| }, |
| { |
| "@type": "cr:FileObject", |
| "@id": "soft_PascalVOC_2012_detection_val.json", |
| "name": "soft_PascalVOC_2012_detection_val.json", |
| "encodingFormat": "application/json", |
| "md5": "84d3750b329158c48e3dbf8a82d37745", |
| "contentSize": "130000000", |
| "description": "Annotations for the Pascal VOC 2012 detection task, validation data (5,823 images)", |
| "contentUrl": "https://huggingface.co/datasets/JKlees/Uncertainty_Aware_Object_Detection/blob/main/soft_PascalVOC_2012_detection_val.json" |
| } |
| ] |
| } |
