Find research datasets worth reusing

opencc-by-4.0Feb 2022View details →

Multi-view rendered YCB dataset for mobile manipulation

This dataset contains different scenarios wherein a mobile robot is approaching a set of YCB objects using both its base and arm motions. There are a total of sixteen sequences with around 100-time steps per sequence. All the sequences were generated using BlenderProc photo-realistic renderer (https://github.com/DLR-RM/BlenderProc). Eight different YCB objects were used. All these objects have a unique 6D pose, while some of the objects also have a single or multiple axes of symmetry. In each sequence maximum of three objects were randomly sampled. In addition, for each sequence, there are five views of the objects from external cameras placed between 2.5-3-5 m facing towards the objects.  This dataset was used for experimental evaluation in the following ICRA 2022 paper: Naik, L., Iversen, T. M., Kramberger, A., Wilm, J., & Krüger, N. (Accepted/In press). Multi-view object pose distribution tracking for pre-grasp planning on mobile robots. In 2022 IEEE International Conference on Robotics and Automation (ICRA) IEEE.   Technical details: In each sequence, the first 5 frames (0-4) contain views from external cameras while frames (5-104) provides a view of the objects visible in the robot camera as it approaches the objects. All the ground truths are provided using the 'coco' annotations format.

opencc-by-4.0Aug 2019View details →

Robot view — Supplementary dataset of experiment videos, IROS 2019 — Self-organized adaptive paths in multi-robot manufacturing: reconfigurable and pattern-independent fibre deployment

This is a supplementary dataset of experiment videos of self-organized multi-robot fibre deployment. Each video is true speed and shows the full respective experiment. These videos show the robot view of each experiment. For a 2-minute summary video of these experiments, refer to: <pre>https://doi.org/10.5281/zenodo.3357187</pre> This supplementary dataset accompanies a conference paper prepared for IEEE IROS 2019. Using multi-robot systems for autonomous construction allows for parallelization and scalability. Swarm construction furthermore exploits robot interactions and collaboration, such that the robot swarm collectively constructs artifacts beyond what a single comparable robot could achieve. Here we present an alternative concept of swarm construction that is distinct because it uses continuous building material. Our approach is unique in its use of braiding techniques for construction. We deploy fibres that potentially allow for structures that are not possible with building blocks. To achieve maximal scalability we restrict ourselves to a decentralized approach. The main challenges are the local coordination of the robot teams, self-organized task allocation, and the dynamic reconfiguration of the braiding scheme at runtime. We successfully validate our approach in multi-robot experiments that show both braiding and branching of the braid. In addition, we show options for implementing an open system—that is robots can join and leave the braiding process on the fly.

opencc-by-4.0Aug 2019View details →

Fibre view — Supplementary dataset of experiment videos, IROS 2019 — Self-organized adaptive paths in multi-robot manufacturing: reconfigurable and pattern-independent fibre deployment

This is a supplementary dataset of experiment videos of self-organized multi-robot fibre deployment. Each video is true speed and shows the full respective experiment. These videos show the fibre view of each experiment. For a 2-minute summary video of these experiments, refer to: <pre>https://doi.org/10.5281/zenodo.3357187</pre> This supplementary dataset accompanies a conference paper prepared for IEEE IROS 2019. Using multi-robot systems for autonomous construction allows for parallelization and scalability. Swarm construction furthermore exploits robot interactions and collaboration, such that the robot swarm collectively constructs artifacts beyond what a single comparable robot could achieve. Here we present an alternative concept of swarm construction that is distinct because it uses continuous building material. Our approach is unique in its use of braiding techniques for construction. We deploy fibres that potentially allow for structures that are not possible with building blocks. To achieve maximal scalability we restrict ourselves to a decentralized approach. The main challenges are the local coordination of the robot teams, self-organized task allocation, and the dynamic reconfiguration of the braiding scheme at runtime. We successfully validate our approach in multi-robot experiments that show both braiding and branching of the braid. In addition, we show options for implementing an open system—that is robots can join and leave the braiding process on the fly.

opencc-by-4.0Oct 2024View details →

Exploring multi-camera views from user-generated sports videos

<div> The proliferation of mobile devices with video recording capabilities has revolutionized the creation, sharing, and consumption of audiovisual content, turning user-generated video (UGV) platforms into major data sources.  <div> <div> Despite this growth, there is a notable gap in the availability of public datasets featuring multi-angle recordings of sports events captured by various mobile cameras. This led to the creation of the MUVY Dataset, with the name stemming from Multiview User-generated Videos from YouTube. <div> <div> The dataset offers a diverse collection of sports videos from multiple perspectives, without restrictions on video size. In its first version, it covers sports like, American football, artistic gymnastics, athletics, basketball, tennis, and cricket. <div> <div> <div> The dataset addresses common challenges in user-generated videos, such as shaking, occlusions, blurring, and abrupt movements. Each video is accompanied by metadata including camera identification, YouTube URLs, extracted frames, and object annotations. </div> </div> </div> </div> </div> </div> </div> </div>

opencc-by-4.0Mar 2023View details →

Multi-view emotional expressions dataset

Multi-view emotional expressions dataset (MEED) using 2D pose estimation.

dryad36/100

MVCNN++: CAD model shape classification and retrieval using multi-view convolutional neural networks

Deep neural networks have shown promising success towards the classification and retrieval tasks for images and text data. While there have been several implementations of deep networks in the area of computer graphics, these algorithms do not translate easily across different datasets, especially for shapes used in product design and manufacturing domain. Unlike datasets used in the 3D shape classification and retrieval in the computer graphics domain, engineering level description of 3D models do not yield themselves to neat distinct classes. The current study looks at an improved form of the 3D shape deep learning algorithm for classification and retrieval through the use of techniques such as relaxed classification, use of prime angled camera angles for capturing feature detail and transfer learning for reducing the amount of data and processing time needed to train shape recognition algorithms. The proposed algorithm (MVCNN++) builds on top of multi-view convolutional neural network (MVCNN) algorithm, improving its efficacy for manufacturing part classification by enabling use of part metadata, yielding an improvement of almost 6% over the original version. With the explosive growth of 3D product models available in publicly available repositories, search and discovery of relevant models is critical to democratizing access to design models.

opencc-zeroAug 2020View details →

opencc-zeroSep 2015View details →

Registration of multi-view echocardiography sequences using a subspace similarity measure

Data employed for the validation of the PCA-based similarity metric proposed in "Registration of multi-view echocardiography sequences using a subspace similarity measure." Peressutti et al. (under review). Data consists of echocardiography sequences of the Left ventricle of four volunteers (vol_A-vol_D) from different acoustic windows (aw_1-aw_5). The image format is metadata. For each subject, the ground-truth rigid transformations that aligns each sequence to all the others is provided. Such transformations are provided by optically tracking the position of the ultrasound imaging probe.   For more detailed information regarding the datasets, please refer to the paper. Python code for testing the proposed method can be downloaded from (https://github.com/devisperessutti/Python.git), while MATLAB code can be downloaded from  (https://github.com/gomezalberto/Matlab.git).

opencc-by-4.0Apr 2022View details →

Supplementary material: Multi-wavelength view of the close-by GRB~190829A sheds light on gamma-ray burst physics

This repository contains supplementary data regarding the article "Multi-wavelength view of the close-by GRB~190829A sheds light on gamma-ray burst physics" published by the Astrophysical Journal Letters. In particular, the repository contains: <ul> <li>Markov Chain Monte Carlo samples for both the afterglow modelling and the circular gaussian fits to VLBI data</li> <li>clean radio images, residuals and UV coverage plots for all our VLBI epochs</li> </ul> Data formats should be self-explanatory. Do not hesitate to contact us at omsharan.salafia@gmail.com for any question.

opencc-by-nc-sa-4.0Jun 2024View details →

VHAKG: Multi-modal Knowledge Graphs with Multi-view Videos of Daily Activities

<h2>Outline</h2> <ul> <li>This dataset is a multimodal knowledge graph (MMKG) of daily activity videos.</li> <li>This dataset integrates a KG with embedded multi-view videos created by <a href="https://github.com/aistairc/virtualhome_aist">VirtualHome-AIST</a>, an extended version of the VirtualHome simulator, and an event-centric KG generated by <a href="https://github.com/aistairc/virtualhome2kg">VirtualHome2KG</a>.</li> <li>We named this dataset VHAKG (VirtualHome-AIST-KG).</li> </ul> <h2>Details</h2> <ul> <li>VHAKG describes 2D bounding boxes of objects every five frames, compositional activities, primitive actions, target objects, object states, 3D bounding boxes, and their time-series changes.</li> <li>The videos are encoded in base64 and embedded as a literal value.</li> <li>VHAKG consists of 706 daily activity scenarios (e.g., clean desk, cook fried bread, and relax on sofa) and 3,530 videos captured by five synchronized cameras per scenario.</li> <li>The file format is <a href="https://www.w3.org/RDF/">RDF</a> (<a href="https://en.wikipedia.org/wiki/Turtle_(syntax)">Turtle</a>), which can be loaded into various <a href="https://en.wikipedia.org/wiki/Triplestore">Triplestores</a>.</li> <li>VHAKG's vocabularies are defined as an ontology and can be found in vh2kg_schema_v2.0.0.ttl.</li> </ul> <h2>Contents</h2> <ul> <li>vh2kg_video_base64.tar.gz <ul> <li>{activity name}{scene}_{camera}_2dbbox.ttl: KG with video embedded in base64 format, including 2D bounding box data every 5 frames. <ul> <li>To learn more about {scene}, check <a href="https://github.com/xavierpuigf/virtualhome/tree/v2.2.0/simulation#environment">here</a>.</li> <li>To learn more about {camera}, check <a href="https://github.com/aistairc/virtualhome_unity_aist?tab=readme-ov-file#addition-of-new-four-cameras">here</a>.</li> </ul> </li> </ul> </li> <li>vh2kg_event.tar.gz <ul> <li>{activity name}_{scene}.ttl: Event-centric KGs representing video content as sequences of events.</li> <li>vh2kg_schema_v2.0.0.ttl: The ontology file of this dataset.</li> <li>affordance.ttl: The affordance data of objects that were created by crowdsourcing. <ul> <li>Please see Section III.B of <a href="https://doi.org/10.1109/ACCESS.2023.3253807" target="_blank" rel="noopener">this paper</a> for more information.</li> </ul> </li> <li>add_places.ttl: Events in which agents moved from one room to another.</li> </ul> </li> </ul> <h2>Tools</h2> <ul> <li><a href="https://github.com/aistairc/vhakg-tools">A set of tools</a> for searching and extracting videos from VHAKG is available.</li> </ul>

opencc-by-4.0Oct 2024View details →

Multi-view spectral images

Multi-view spectral images took on 07/11/2024 using Confocal 980.

opencc-by-4.0Jul 2021View details →

Shadow Neural Radiance Fields for Multi-View Satellite Photogrammetry - Dataset

Data accompanying the paper Derksen, Dawa, and Dario Izzo. "Shadow Neural Radiance Fields for Multi-view Satellite Photogrammetry." Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. 2021. This dataset contains a subset of the World-View-3 images from the IEEE Data Fusion Competition 2019 - Track 3, downloaded from https://ieee-dataport.org/open-access/data-fusion-contest-2019-dfc2019. It is organized in four folders, one for each study area, following the original area names. Each folder contains a multi-view set of RGB images, cropped to the validation area, and rotated according to the azimuth angle. The folder also contains a Digital Surface Model of the area (DSM) as a one-band .tif file where the values contained in pixels represent the surface altitude in meters. Finally the folder contains a "metadata" file which provides for each image ID the radius (distance from satellite to scene), as well as the viewing and lighting directions (azimuth and elevation). The authors would like to thank the Johns Hopkins University Applied Physics Laboratory and IARPA for providing the data used in this study, and the IEEE GRSS Image Analysis and Data Fusion Technical Committee for organizing the Data Fusion Contest.

opencc-by-4.0Dec 2019View details →

RailEnV-PASMVS: a dataset for multi-view stereopsis training and reconstruction applications

A Perfectly Accurate, Synthetic dataset featuring a virtual railway EnVironment for Multi-View Stereopsis (RailEnV-PASMVS) is presented, consisting of 40 scenes and 79,800 renderings together with ground truth depth maps, extrinsic and intrinsic camera parameters and binary segmentation masks of all the track components and surrounding environment. Every scene is rendered from a set of 3 cameras, each positioned relative to the track for optimal 3D reconstruction of the rail profile. The set of cameras is translated across the 100-meter length of tangent (straight) track to yield a total of 1,995 camera views. Photorealistic lighting of each of the 40 scenes is achieved with the implementation of high-definition, high dynamic range (HDR) environmental textures. Additional variation is introduced in the form of camera focal lengths, random noise for the camera location and rotation parameters and shader modifications of the rail profile. Representative track geometry data is used to generate random and unique vertical alignment data for the rail profile for every scene. This primary, synthetic dataset is augmented by a smaller image collection consisting of 320 manually annotated photographs for improved segmentation performance. The specular rail profile represents the most challenging component for MVS reconstruction algorithms, pipelines and neural network architectures, increasing the ambiguity and complexity of the data distribution. RailEnV-PASMVS represents an application specific dataset for railway engineering, against the backdrop of existing datasets available in the field of computer vision, providing the precision required for novel research applications in the field of transportation engineering.   File descriptions <ul> <li>RailEnV-PASMVS.blend (227 Mb) - Blender file (developed using Blender version 2.8.1) used to generate the dataset. The Blender file packs only one of the HDR environmental textures to use as an example, along with all the other asset textures.</li> <li>RailEnV-PASMVS_sample.png (28 Mb) - A visual collage of 30 scenes, illustrating the variability introduced by using different models, illumination, material properties and camera focal lengths.</li> <li>geometry.zip (2 Mb) - Geometry CSV files used for scenes 01 to 20. The Bezier curve defines the geometry of the rail profile (10 mm intervals).</li> <li>PhysicalDataset.7z (2.0 Gb) - A smaller, secondary dataset of 320 manually annotated photographs of railway environments; only the railway profiles are annotated.</li> <li>01.7z-40.7z (2.0 Gb each) - Archive of every scene (01 through 40).</li> <li>all_list.txt, training_list.txt, validation_list.txt - Text files containing the all the scene names, together with those used for validation (validation_list.txt) and training (training_list.txt), used by MVSNet.</li> <li>index.csv - CSV file provides a convenient reference for all the sample files, linking the corresponding file and relative data path.</li> </ul>   Steps to reproduce The open source Blender software suite (https://www.blender.org/) was used to generate the dataset, with the entire pipeline developed using the exposed Python API interface. The camera trajectory is kept fixed for all 40 scenes, except for small perturbations introduced in the form of random noise to increase the camera variation. The camera intrinsic information was initially exported as a single CSV file (scene.csv) for every scene, from which the camera information files were generated; this includes the focal length (focalLengthmm), image sensor dimensions (pixelDimensionX, pixelDimensionY), position, coordinate vector (vectC) and rotation vector (vectR). The STL model files, as provided in this data repository, were exported directly from Blender, such that the geometry/scenes can be reproduced. The data processing below is written for a Python implementation, transforming the information from Blender's coordinate system into universal rotation (R_world2cv) and translation (T_world2cv) matrices.   <pre><code class="language-python">import numpy as np from scipy.spatial.transform import Rotation as R #The intrinsic matrix K is constructed using the following formulation: focalLengthPixel = focalLengthmm x pixelDimensionX / sensorWidthmm K = [[focalLengthPixel, 0, dimX/2], [0, focalPixel, dimY/2], [0, 0, 1]] #The rotation vector as provided by Blender was first transformed to a rotation matrix: r = R.from_euler('xyz', vectR, degrees=True) matR = r.as_matrix() #Transpose the rotation matrix, to find matrix from the WORLD to BLENDER coordinate system: R_world2bcam = np.transpose(matR) #The matrix describing the transformation from BLENDER to CV/STANDARD coordinates is: R_bcam2cv = np.array([[1, 0, 0], [0, -1, 0], [0, 0, -1]]) #Thus the representation from WORLD to CV/STANDARD coordinates is: R_world2cv = R_bcam2cv.dot(R_world2bcam) #The camera coordinate vector requires a similar transformation moving from BLENDER to WORLD coordinates: T_world2bcam = -1 * R_world2bcam.dot(vectC) T_world2cv = R_bcam2cv.dot(T_world2bcam)</code></pre>   The resulting R_world2cv and T_world2cv matrices are written to the camera information file using exactly the same format as that of <a href="https://github.com/YoYo000/BlendedMVS">BlendedMVS developed by Dr. Yao</a>. The original rotation and translation information can be found by following the process in reverse. Note that additional steps were required to convert from Blender's unique coordinate system to that of OpenCV; this ensures universal compatibility in the way that the camera intrinsic and extrinsic information is provided. Equivalent GPS information is provided (gps.csv), whereby the local coordinate frame is transformed into equivalent GPS information, centered around the <a href="https://www.up.ac.za/eng4">Engineering 4.0 campus, University of Pretoria</a>, South Africa. This information is embedded within the JPG files as EXIF data.

dryad36/100

MVCNN++: CAD model shape classification and retrieval using multi-view convolutional neural networks

Open the record for dataset details and reuse information.

publicAug 2020View details →

opencc-by-4.0Nov 2024View details →

wikidata Dataset for Multi-View Structural Graph Summaries

Open the record for dataset details and reuse information.

opencc-by-4.0Jul 2024View details →

CSS3V Dataset for Multi-View Structural Graph Summaries

Open the record for dataset details and reuse information.

dryad32/100

Data from: A multi-tissue view on telomere dynamics and postnatal growth

Trade-offs between growth and self-maintenance are common in nature, such that early-life effects on growth can generate lasting consequences on survival and longevity. Telomeres – putative biomarkers of self-maintenance – may link early growth with these later phenotypic effects, but evidence for growth-telomere trade-offs is mixed. Null or even positive relationships between growth and telomeres may be driven by heterogeneity in resource availability or invariable allocation towards telomere maintenance within a population. We used nestling tree swallows (Tachycineta bicolor) to assess the directionality and timing of relationships between growth and telomere length in several tissues. We focused on two important phases of growth: first, the peak of postnatal growth occurring around 6-days old when nestlings grow by ~33% in a single day, and subsequently, the later phase of growth occurring as body mass plateaus near adult size at 12-days old. We quantified telomere attrition in blood during postnatal growth, as well as telomere length in the blood, brain, adrenals, and liver at 12-days old. Growth was unrelated to telomere length in the liver and telomere dynamics in blood. However, brain telomere length was positively correlated with peak growth, and adrenal telomere length was positively related to later growth, particularly for chicks that had experienced a temporary stressor. These observations suggest that variation in resource availability may mask trade-offs, generating positive correlations between growth and telomere length at the population level. They also provide insights into complex relationships between growth and self-maintenance that can be revealed by looking in multiple tissues.

opencc-zeroJan 2022View details →

openother-ncFeb 2022View details →

MM-Office Dataset: multi-view and multi-modal dataset in an office environment

MM-office is a multi-view and multi-modal dataset in an office environment (MM-Office) that records events, e.g., 'enter' to the office room, 'sit down' on the chair, and 'take out' something from a shelf, in the room assuming the daily work. These events are recorded simultaneously using eight non-directional microphones and four cameras. The audio and video clips are divided into scenes, each about 30 to 90 seconds. The amount of data was 880 clips per point and sensor. The labels available for training are given as multi-labels that indicate which each clip contains what event. Only the test data is annotated with a strong label containing the onset/offset time of each event. License: see the file named LICENSE.pdf Further information is available at [1] and Github: https://github.com/nttrd-mdlab/mm-office [1] Masahiro Yasuda, Yasunori Ohishi, Shoichiro Saito, Noboru Harada “Multi-view and Multi-modal Event Detection Utilizing Transformer-based Multi-sensor fusion,” in IEEE Int. Conf. Acoust. Speech Signal Process. (ICASSP), 2022.