Find research datasets worth reusing
Search datasets from major research repositories and use ShareScore to quickly assess how well each record supports discovery, access, and reuse.
1,506
datasets available to search
ShareScore release 0.9.0
Dataset results
1,506 results for “objects”
Corpus Nummorum - Object Detection Coin Dataset
<p>This Object Detection dataset is a collection of ancient coin images from three different sources: the <a href="https://www.corpus-nummorum.eu/">Corpus Nummorum (CN)</a> project, the <a href="https://ikmk.smb.museum/home?lang=en">Münzkabinett Berlin</a> and the <a href="https://www.bnf.fr/fr/departement-monnaies-medailles-antiques">Bibliothèque nationale de France, Département des Monnaies, médailles et antiques</a>. It covers Greek and Roman coins from ancient Thrace, Moesia Inferior, Troad and Mysia. This is a selection of the coins published on the <a href="https://www.corpus-nummorum.eu/search/coins">CN</a> portal (due to copyrights). </p> <p>This dataset contains 506 different classes with about 179.000 coin images (approx. 29.000 unique coins). The classes come from four different categories: persons, objects, animals and plants. The coin images were assigned to the classes using our <a href="https://github.com/Frankfurt-BigDataLab/NLP-on-multilingual-coin-datasets">NLP pipeline</a>. For this purpose, our Named Entity Recognition and Relation Extraction were performed on every coin's description (separated into obverse and reverse). Each coin image assigned to this description was then <span>copied to the folder of the predicted classes</span>. A coin image can therefore also be assigned to different classes. The file name contains both the coin id and the coin type of the <a href="https://www.corpus-nummorum.eu/">CN database</a>. Whether the image belongs to a coin obverse or reverse can be recognized by the suffix obv or rev. An "sources" csv file holds the sources for every image. Due to copyrights the image size is limited to 299*299 pixels. However, this should be sufficient for most ML approaches.</p> <p>Due to the numerically different occurrences of the individual entities, the data set is not balanced. In addition, a class can contain very different representations of the same entity. Therefore, some classes can be difficult to train. Unfortunately, we cannot provide any annotations for the data set.</p> <p>During the summer semester 2024, we held the "Data Challenge" event at our Department of Computer Science at the Goethe-University. Our students could choose between the Object Detection dataset and a Natural Language dataset as their challenge. One team opted for the Object Detection challenge. We gave them this dataset with the task to use to try out their own ideas. Here are their results:</p> <p><a href="https://github.com/PatrickMelnic/DataChallenges_ObjectDetection/">Multilabel Classification as Backbone for Object Detection</a></p> <p> </p> <p>Now we would like to invite you to try out your own ideas and models on our coin data.</p> <p>If you have any questions or suggestions, please, feel free to contact us. </p>
Dangerous Items Dataset for 5-Class Object Detection (Pascal VOC annotation)
<p>This repository contains the data from the manuscript "Dangerous Items Detection in Surveillance Camera Images Using Faster R-CNN". It contains 4,000 images representing 5 classes of objects: baseball bat, gun, knife, machete and rifle (<a href="https://drive.google.com/file/d/1aG30Hpupctnp8ywhAprz313jmYbFnZk3/view?usp=drive_link" target="_blank" rel="noopener">https://drive.google.com/file/d/1aG30Hpupctnp8ywhAprz313jmYbFnZk3/view?usp=drive_link</a>). All images were scaled so that the smaller side is no shorter than 600 pixels and the larger one is no longer than 1,000 pixels. The full set was randomly divided into training and testing parts (75% and 25% of the full set respectively). As a result, the training part contains 3,000 images, and the testing part – 1,000 images. Both parts are balanced, that is, they contain similar number of objects to be detected. Images were annotated using bounding boxes in the Pascal VOC format, according to which the description about each image is included in the corresponding XML file.</p> <p>Using this dataset please cite:<br>Omiotek, Z. (2025). Dangerous items’ detection in surveillance camera images using Faster R‑CNN. Przegląd Elektrotechniczny, 101(5), 156-168. https://www.red.pe.org.pl/articles/2025/5/36.pdf</p>
Popular Animals Dataset for 6-Class Object Detection (Pascal VOC annotation)
<p>This dataset contains images used in the monograph titled <em>Zastosowanie wybranych metod uczenia głębokiego w wizji komputerowej</em> (Application of Selected Deep Learning Methods in Computer Vision) to build the Faster R-CNN model. The full collection consists of 600 image files showing 6 classes of objects: <em>kot</em> (cat), <em>krowa</em> (cow), <em>pies</em> (dog), <em>koń</em> (horse), <em>człowiek</em> (human), <em>owca</em> (sheep) (<a title="Popular Animals Dataset" href="https://drive.google.com/file/d/1g3O1Qq2YqmCb8WSMDHaoFJPoHe6yujlQ/view?usp=drive_link" target="_blank" rel="noopener">https://drive.google.com/file/d/1g3O1Qq2YqmCb8WSMDHaoFJPoHe6yujlQ/view?usp=drive_link</a>). All images were scaled so that the smaller side is no shorter than 600 pixels and the larger side is no longer than 1000 pixels. The set was randomly divided into a training part (75% of the full set) and a test part (25% of the full set). As a result, the training part contains 450 files, and the test part - 150. Both parts are balanced - they contain a similar number of detected objects. The images are labeled with bounding boxes in the Pacal VOC format, according to which the description of each file is contained in an XML file with the same name. The dataset can be used to build models for object detection.</p>
BadODD: Bangladeshi Autonomous Vehicle Object Detection Dataset
<p>The dataset covers the following 9 districts in Bangladesh: Sylhet, Dhaka, Rajshahi, Mymensingh, Maowa, Chittagong, Sirajganj, Sherpur, and Khulna. Participants will encounter a wide range of road types, including towns, expressways, highways, and village roads. This diversity in locations aims to challenge algorithms to perform well across various driving contexts commonly encountered on Bangladesh roads.</p> <p> </p>
Deep-sea observatories images labeled by citizen for object detection algorithms
<div>All information is available from the original publication page: <a href="https://doi.org/10.17882/101899">https://doi.org/10.17882/101899</a>.</div>
Wide-angle simulated artificial vision enhances spatial navigation and object interaction in a naturalistic environment
<p>This folder contains the dataset and the processing code generated in the article 'Wide-angle simulated artificial vision enhances spatial navigation and object interaction in a naturalistic environment' by Hinrichs et al. </p> <p> </p>
RBOD: An annotated satellite imagery dataset for automated river barrier object detection
<h3>Introduction:</h3> <p>Millions of river barriers have been constructed worldwide for flood control, hydropower generation, and agricultural irrigation. The lack of comprehensive records on the locations and types of river barriers, particularly small barriers such as weirs, limits our ability to assess their societal and environmental impacts. Integrating satellite imagery with object detection algorithms holds promise for the automatic identification of river barriers on a global scale. However, achieving this objective requires high-quality image datasets for algorithm training and testing. Hence, this study presents a large-scale dataset named the River Barrier Object Detection (RBOD), making it the first publicly available dataset specifically for river barrier object detection.</p> <p>The RBOD dataset comprises 4,872 high-resolution satellite images and 11,741 meticulously annotated oriented bounding boxes (OBBs). In this dataset, river barriers can be classified into five classes: dams, groynes, locks, sluices, and weirs. The effectiveness of the RBOD dataset was validated using five typical object detection algorithms, namely YOLOV8-OBB, Oriented R-CNN, Rotated Faster R-CNN, R3Det, and Rotated RetinaNet, which provide performance benchmarks for future applications.</p> <h3><span>Usage Notes:</span></h3> <p>The RBOD dataset consists of three folders (namely, '<em>images</em>', '<em>labels_voc</em>', and '<em>labels_yolo</em>') and a .txt file named '<em>class</em>':</p> <p>·'<em>images</em>' folder - contains 4872 satellite images (.jpg).</p> <p>·'<em>labels_voc</em>' folder - contains 11,741 .xml files for annotations in PASCAL VOC format. In these .xml files, the position of OBB is represented as (cx, cy, width, height, angle), where 'cx' and 'cy' denote the center coordinates, 'width' and 'height' are the lengths along the x- and y-axes, and 'angle' is the clockwise rotation angle relative to the x-axis.</p> <p>·'<em>labels_yolo</em>' folder - contains 11,741 .txt files for annotations in YOLO format. In these .txt files, the OBB is represented as (class_index, x1, y1, x2, y2, x3, y3, x4, y4), where ‘class_index’ denotes the target category, and ‘x1, y1, x2, y2, x3, y3, x4, y4’ are the normalized coordinates of the four corners of the bounding box.</p> <p>·'<em>class</em>' txt file - record the classifications of river barriers and their indices, which correspond to the ‘class_index’.</p> <p>Note that each folder splits into three subfolders: train (70%), test (20%), and val (10%).</p>
Comparison and assessment of different object-based classifications using machine learning algorithms and UAVs multispectral imagery in the framework of precision agriculture
<p>Supplementary material of the paper</p>
Salient objects dominate the central fixation bias when orienting towards images
<p>Data and stimulus material from:</p> <p>Wolf, C., & Lappe, M. (2021). Salient objects dominate the central fixation bias when orienting towards images. Journal of Vision</p> <p>For questions please contact chr.wolf[at]wwu.de</p>
Puram et al HNSCC data SingleCellExperiment and seurat object for MultiNicheNet
<p>Puram et al HNSCC data SingleCellExperiment and seurat object for MultiNicheNet</p>
Figure 3 in Amplified fragment length polymorphisms, the evolution of the land snail genus Theba (Stylommatophora: Helicidae), and an objective approach for relating fossils to internal nodes of a phylogenetic tree using geometric morphometrics
Figure 3. Thin plate splines illustrating shape changes between selected nodes of the tree in Figure 4 based on weighted branch lengths.
Figure 4 in Amplified fragment length polymorphisms, the evolution of the land snail genus Theba (Stylommatophora: Helicidae), and an objective approach for relating fossils to internal nodes of a phylogenetic tree using geometric morphometrics
Figure 4. Reconstruction of shell shape and size based on weighted (above branches) and unweighted (below branches) branch lengths. The inset shows the tree shape based on COI sequence data evolved into the AFLP tree topology. Node numbers are in italic; size is expressed as centroid size; the colour of the centroid size values indicates shape changes.
Data objects
<p>The list of the data objects in the repository.</p>
FIGURES 8–9 in A sketch-based system for modeling 3D objects: Applications to taxonomy
FIGURES 8–9. Examples of usages of our system. Fig. 8. Sketching, visualization and reconstruction of an open contour. (a) Haken's gordian knot (adapted from Fish & Lisitsa 2014); (b) our sketch traced over the original drawing (adapted from Burton
FIGURE 14 in A sketch-based system for modeling 3D objects: Applications to taxonomy
FIGURE 14. Male terminalia of Chrysopilus (Diptera, Rhagionidae): (a) Epandrium, dorsal view, of C. phaeopterus (from Santos & Amorim 2007); (b) our sketch traced over the original illustration; and (c) The 3D resulting model.
FIGURE 1 in A sketch-based system for modeling 3D objects: Applications to taxonomy
FIGURE 1. Categories supported as inputs for the reconstruction of a chromosome. Input strokes belong to three categories: (a) open contours and their respective 3D reconstruction; (b) closed contours and 3D surface reconstruction; (c) stripes sketched using our system and interactions and their 3D reconstruction; and (d) final 3D model combining the three categories.
FIGURE 16 in A sketch-based system for modeling 3D objects: Applications to taxonomy
FIGURE 16. Head of Austroleptis papaveroi (Diptera, Austroleptidae): (a) photograph using a microscope (from Fachin et al. 2018); (b) our closed contour selection based on the photograph; (c) colored layers; and (d)–(f) resulting 3D model.
FIGURES 2–7 in A sketch-based system for modeling 3D objects: Applications to taxonomy
FIGURES 2–7. Creation, visual enhancement and 3D reconstruction. Fig. 2. Strokes: (a) raw stroke; (b) default smoothing applied three times; (c) smoothing applied nine times; (d) oversketching through drawing a new curve near the original one; and (e) replacing the current sketch with the new curve. Fig. 3. Selection operation: (a) cross selection (green) for a closed contour; (b) brush selection (red) by painting the segments belonging to the same contour. Fig. 4. Reconstruction process of a 3D stripe: (a) lines drawn with twist annotation using inverted colors for stripe sides; (b) automatic creation of stripe twists; (c) 3D parametric reconstruction of the stripe. Fig. 5. Sketch inference: (a) triangles created combining points; (b) "headlight beam" effect created by painting triangles according to their angle. Fig. 6. Different depth effects: (a) contours using gray-scale to depict different layers; (b) halo effect; (c) hatching effect; (d) layer coloring from blue to yellow. Fig. 7. Smoothing operations of an open contour. (a) one time; (b) ten times; (c) fifty times; (d) one hundred times; and (e) two hundred times.
FIGURE 15 in A sketch-based system for modeling 3D objects: Applications to taxonomy
FIGURE 15. Male terminalia of Chrysopilus (Diptera, Rhagionidae): (a) hypopygium, ventral view, of C. balbii (from Santos & Amorim 2017); (b)–(c) our sketch traced over the original illustration; and (d)–(f) The resulting 3D model.
FIGURE 13 in A sketch-based system for modeling 3D objects: Applications to taxonomy
FIGURE 13. Female terminalia of Austroleptis papaveroi (Diptera, Austroleptidae): (a) photograph using a microscope; (b) drawing used as input for our 3D reconstruction (both from Fachin et al. 2018); (c)–(d) our closed contour selection traced from the adapted illustration; and (e)–(f) the resulting 3D model.
ScienceDex guides
Understand access before you commit
These curated guides explain access requirements, typical timelines, costs, and reuse considerations for widely used research datasets.
Allen Brain Atlas
Allen Brain Atlas is an Allen Institute collection of brain map atlases, datasets, APIs, and analysis tools covering mouse, human, and non-human primate brain resources.
Annotated Behaviour and Observability Dataset (ABODe)
ABODe is a University of Edinburgh DataShare dataset for behavior classification in group-housed mice using home-cage video, identities, bounding boxes, ground-plate positions, and annotator labels.
DANDI Archive for NWB datasets
DANDI is a BRAIN Initiative archive for publishing and sharing neurophysiology data, including electrophysiology, optophysiology, and behavioral data packaged as NWB and related standards.
International Brain Laboratory public data
The International Brain Laboratory public data releases expose standardized mouse decision-making experiments, including Neuropixels recordings, widefield calcium imaging, behavior, and session metadata accessed through the ONE API.
OpenNeuro
OpenNeuro is a free, open platform for sharing neuroimaging datasets, with public search, dataset pages, and download paths for web, S3, DataLad, and the OpenNeuro CLI.