Find research datasets worth reusing

opencc-by-3.0-usJul 2022View details →

Dataset for a physical model characterizing visualization of the cervix during pelvic exams

This dataset accompanies our manuscript draft: A physical model for improving visualization of the cervix  during pelvic exams: A steppingstone towards reducing  disparities in women's health. Manuscript Draft Abstract Pelvic exams are frequently complicated by collapse of the lateral vaginal walls, obstructing the physician’s view of the cervix. A commonly utilized method in the clinical setting, passed down from mentors to trainees, is repurposing either a condom or a glove as a sheath placed over the speculum blades to retract the lateral vaginal walls during the exam. Despite their regular use in clinical practice, little research has been done comparing the relative efficacy of these methods. Better visualization of the cervix can benefit patients by decreasing examination-related discomfort, aiding in cancer screening, and preventing the need to move the examination to the operating room under general anesthesia. This study presents a physical model that simulates vaginal pressure being exerted around a speculum. Using it, we then compare the efficacy of different condom types, glove materials, glove sizes, and methods of application onto the speculum. The results showed that condoms provided minimal lateral wall retraction, while vinyl-material gloves with the speculum placed into the third finger had the best lateral wall retraction. However, the nitrile-material gloves are overall preferred over the vinyl gloves as they provided adequate lateral wall retraction without applying a significant vertical compressive effect on the speculum, and thus had overall better cervical visualization. Glove size had minimal impact. This study serves as a guide for clinicians as they use tools commonly found in a clinical setting to perform difficult pelvic exams. We recommend that clinicians consider the use of a nitrile glove as a sheath around a speculum. Additionally, this study demonstrates proof-of-concept of a physical model that can quantitatively describe different materials on their ability to improve cervical visualization. This model can be used in future research with more speculum and material combinations, including with materials custom-designed materials for this purpose.

opencc-by-4.0Jul 2018View details →

Data accompanying the master thesis: A neuronal model for visually evoked startle responses in schooling fish

This dataset contains data that was generated and analyzed for the master thesis "A neuronal model for visually evoked startle responses". All related material, including analysis code, of the master thesis can be found at https://github.com/awakenting/master-thesis.

opencc-by-4.0Dec 2019View details →

Visual and inertial data for validation of gliding models of ornithopters

This dataset contains data from different gliding flights with an ornithopter in low wind conditions. For each experiment, the inertial information is provided. Additionally, the flights have been recorded from three different points of view to track and triangulate its position. The videos are provided and the position of the camera has been determined using a Leica Total Station system with submillimeter accuracy. A sample of the 2D track of the ornithopter is provided for each video and experiment. The tracking along the three cameras are synchronized.   ------Camera Pose structure------   Three cameras with four points: three to measure orientation and the last one the lens position. The last two points are the measured fall. Camera 1 -> top left, bottom left and top right. Camera 2 -> top left, top rigth and bottom right. Camera 3 -> top left, bottom left and top right. Then there are 14 rows. The pattern is: Point1, Point2, Point3 and Lens Position. ------IMU structure------ time, quaternion w, quaternion x, quaternion y, quaternion z, accelerometer x, accelerometer y, accelerometer z, Gyroscope x, Gyroscope y, Gyroscope z, magnetometer x ,magnetometer y ,magnetometer z units: time->ms, accelerometer->g, gyroscope->ยบ/s

opencc-by-4.0Mar 2023View details →

Supplementary data for the paper "Visual integration of omics data to improve 3D models of fungal chromosomes"

<ul> <li>13 parameter files (*.YML) used by the 3DGB workflow to produce models of 3D genomes.</li> <li>13 3D genomes structures (*.PDB).</li> <li>4 animated GIF of representative structures.</li> <li>1 XLSX file that lists raw (Hi-C and ChIP-seq) data used in this study and the associated analysis.</li> </ul>

openmit-licenseJul 2023View details →

lilGym: Natural Language Visual Reasoning with Reinforcement Learning, model files

Baselines models for the paper <a href="https://lil.nlp.cornell.edu/lilgym">lilGym: Natural Language Visual Reasoning with Reinforcement Learning</a>.

dryad40/100

Asymmetry in kinematic generalization between visual and passive lead-in movements are consistent with a forward model in the sensorimotor system

In our daily life we often make complex actions comprised of linked movements, such as reaching for a cup of coffee and bringing it to our mouth to drink. Recent work has highlighted the role of such linked movements in the formation of independent motor memories, affecting the learning rate and ability to learn opposing force fields. In these studies, distinct prior movements (lead-in movements) allow adaptation of opposing dynamics on the following movement. Purely visual or purely passive lead-in movements exhibit different angular generalization functions of this motor memory as the lead-in movements are modified, suggesting different neural representations. However, we currently have no understanding of how different movement kinematics (distance, speed or duration) affect this recall process and the formation of independent motor memories. Here we investigate such kinematic generalization for both passive and visual lead-in movements to probe their individual characteristics. After participants adapted to opposing force fields using training lead-in movements, the lead-in kinematics were modified on random trials to test generalization. For both visual and passive modalities, recalled compensation was sensitive to lead-in duration and peak speed, falling off away from the training condition. However, little reduction in force was found with increasing lead-in distance. Interestingly, asymmetric transfer between lead-in movement modalities was also observed, with partial transfer from passive to visual, but very little vice versa. Overall these tuning effects were stronger for passive compared to visual lead-ins demonstrating the difference in these sensory inputs in regulating motor memories. Our results suggest these effects are a consequence of state estimation, with differences across modalities reflecting their different levels of sensory uncertainty arising as a consequence of dissimilar feedback delays.

opencc-zeroFeb 2020View details →

opencc-by-4.0Nov 2023View details →

Training and test data, plus saved models for the upcoming paper `Top-down perceptual inference shaping the activity of early visual cortex'

Each .pkl file contains a training or test dataset in the form of a Python dictionary (generated with Python 3.8.5) with the following fields:<ul><li>'train_images': 640,000 float32 images used for model training. These are 40px images that contain 1600 pixel intensities each.</li><li>'train_labels': float32 labels for each image in 'train_images'. All natural images are labeled with 0.0. Texture images are labeled with 0.0, 1,0, 2.0, 3.0, or 4.0, according to their texture family.</li><li>'test_images': 64,000 float32 images used for model testing. These are 40px images that contain 1600 pixel intensities each.</li><li>'test_labels': float32 labels for each image in 'test_images'. All natural images are labeled with 0.0. Texture images are labeled with 0.0, 1,0, 2.0, 3.0, or 4.0, according to their texture family.</li></ul>The .zip file contains a saved model snapshot and various intermediate evaluative data. Details on these are coming soon.

opencc-by-4.0Dec 2023View details →

Supplementary material for "Exploring Conceptual Data Modeling Processes: Insights from Clustering and Visualizing Modeling Sequences"

This material supplements the following conference publication: Winkler, Rosenthal, Strecker (2024). "Exploring Conceptual Data Modeling Processes: Insights from Clustering and Visualizing Modeling Sequences". Modellierung 2024.

opencc-by-4.0Jul 2022View details →

Prediction and Visualization of Human Transmembrane Proteins using AlphaFold and Protein Language Models

Description: TMvis ("TMvis496.tar.gz") is a dataset containing 496 3D-structures of predicted human transmembrane proteins (TMP) and their predicted membrane embedding. The method TMbed [1], based on the protein language model ProtT5 [2] predicted 4.967 TMP for the human proteome (20,375 proteins, UniProt [3] version April 2022; excluding TITIN_HUMAN due to length). For these proteins, we obtained AlphaFold [4] structures from AlphaFoldDB [5] with an average per-residue confidence score (pLDDT) of more than 90%. This resulted in the 496 proteins of TMvis, as can be found in "TMvis496.fasta". The membrane embedding was predicted using the methods ANVIL [6], PPM3 [7], and per-residue TMbed predictions. As the three methods are based on different approaches, we decided to publish results for all. The figure “TMvis_project_overview.png” provides a graphical overview for each step described above. TMvis Folder Structure: TMvis is separated into “alpha” containing predicted alpha-helical TMPs, and “beta” containing predicted beta-barrel TMPs. Within these folders, each protein is assigned one folder, identifiable by the respective unique UniProt ID. Each protein folder consists of: - “UniprotID.fasta” with UniProt ID, sequence, TMbed per-residue prediction - “AF-UniprotID-F1-model_v2.pdb” with the AlphaFold structure - “AF-UniprotID-F1-model_v2.cif” with the AlphaFold structure - “AF-UniprotID-F1-model_v2_ANVIL.pdb” with predicted ANVIL membrane embedding - “AF-UniprotID-F1-model_v2_ppm.pdb” predicted PPM3 membrane embedding TMvis                            |                                     ├── alpha                                  │   │                                  │   ├── A0A087X1C5                                  │   │   ├── A0A087X1C5.fasta                                  │   │   ├── AF-A0A087X1C5-F1-model_v2.pdb                                  │   │   ├── AF-A0A087X1C5-F1-model_v2.cif                                  │   │   ├── AF-A0A087X1C5-F1-model_v2_ANVIL.pdb                                  │   │   └── AF-A0A087X1C5-F1-model_v2_ppm.PDB                                  │   └── ...                                     └── beta                                      └── P45880 TMvis visualization: The 3D-visualization of every protein in the dataset TMvis can be easily accessed using the Jupyter Notebook “TMvis.ipynb”. It contains detailed descriptions the different membrane prediction tools ANVIL, PPM3, and TMbed as well as the respective code. Additionally, it allows to visualize the per-residue confidence scores (pLDDT) of AlphaFold. —————————————————————————————————————————————————————————————————————————— References: [1] TMbed - TMbed Bernhofer, Michael, and Burkhard Rost. 2022. “TMbed – Transmembrane Proteins Predicted through Language Model Embeddings.” bioRxiv. [2] ProtT5 - A. Elnaggar et al., "ProtTrans: Towards Cracking the Language of Lifes Code Through Self-Supervised Deep Learning and High Performance Computing," in IEEE Transactions on Pattern Analysis and Machine Intelligence, doi: 10.1109/TPAMI.2021.3095381. [3] UniProt - UniProt Consortium (2021). UniProt: the universal protein knowledgebase in 2021. Nucleic acids research, 49(D1), D480–D489. [4] AlphaFold - AlphaFold Jumper, John, Richard Evans, Alexander Pritzel, Tim Green, Michael Figurnov, Olaf Ronneberger, Kathryn Tunyasuvunakool, et al. 2021. “Highly Accurate Protein Structure Prediction with AlphaFold.” Nature 596 (7873): 583–89. [5] Alphafold DB - Varadi, Mihaly, Stephen Anyango, Mandar Deshpande, Sreenath Nair, Cindy Natassia, Galabina Yordanova, David Yuan, et al. 2022. “AlphaFold Protein Structure Database: Massively Expanding the Structural Coverage of Protein-Sequence Space with High-Accuracy Models.” Nucleic Acids Research 50 (D1): D439–44. [6] ANVIL - ANVIL Postic, Guillaume, Yassine Ghouzam, Vincent Guiraud, and Jean-Christophe Gelly. 2016. “Membrane Positioning for High- and Low-Resolution Protein Structures through a Binary Classification Approach.” Protein Engineering, Design & Selection: PEDS 29 (3): 87–91. [7] PPM3 - PPM3 Lomize, Mikhail A., Irina D. Pogozheva, Hyeon Joo, Henry I. Mosberg, and Andrei L. Lomize. 2012. “OPM Database and PPM Web Server: Resources for Positioning of Proteins in Membranes.” Nucleic Acids Research 40 (Database issue): D370–76. —————————————————————————————————————————————————————————————————————————— License: This work is licensed under a Creative Commons Attribution 4.0 International License (CC-BY 4.0).

opencc-by-4.0Jun 2010View details →

Figure 3. Data visualization-DATA MINING LEARNING MODELS AND ALGORITHMS ON A SCADA SYSTEM DATA REPOSITORY

Data visualization is also a very useful technique because it helps to deter- mine the di±culty of the learning problem. We visualized with Weka single attributes (1-d) and pairs of attributes (2-d). The ¯gure 3 shows the variation of the temperature in time.

opencc-by-4.0Jan 2016View details →

Figure 1. Visual and synthetic representation of the modelling process in the software industry-The Fundamentals Regarding the Usage of the Concept of Interface for the Modeling of the Software Artefacts

The experience that is accumulated regarding the modelling paradigms in the software engineering is impressive. Thus, the software engineering recognizes modelling paradigms like object orientation, aspect orientation, component orientation, service orientation, agent orientation. In one form or another, these paradigms prove their ex- cellence in certain types of IT projects. At the same time, these paradigms reveal their objective limits when they are used to engineer the real world software systems. Every modelling paradigm represents, in fact, a modality to represent the real world using a specific formal framework. The specificity of the formal framework is defined from both a syntactic and semantic perspective. The  formal syntactic framework of a paradigm refers to the  concepts that are  used  by the  paradigm in  order to represent the  real  world,  but  also  to the recommended principles that allow  for these concepts to interact in a correct and  efficient manner. Both the concepts and the principles benefit from a formal representation that ultimately favours communication as a secondary modelling lever inside the IT projects. Every syntactic artefact of a paradigm can be associated with a certain real world semantics, which it abstracts. As a consequence, considering that the real world continuously enhances its semantic potential, the syntactic constructs that are favoured by the paradigm may become problematic.

Prototyping 3D Virtual Learning Environments with X3D-based Content and Visualization Tools-Figure 10. Room model generated with Autodesk 123D Catch - the 3D model (screen capture from GLC Player)

Structure from motion was used for rapid modeling of a small room with all its objects. Two files were generated, a Wavefront obj and mtl (corresponding to the texture). The 3D model was post-processed with MeshLab, during which several filters were applied to clean up the model. The mesh model was also connected with the scanned model, by choosing at least 4 connection points. The 2D and 3D results are shown in Figures 9, 10. A post-processing could also be performed using the Autodesk 123D Catch web application.

Prototyping 3D Virtual Learning Environments with X3D-based Content and Visualization Tools-Figure 2. The model of the 3D virtual campus - details from the building interior (3D modeling by Marius Hodea)

The processing workflow for 3D modeling and design for a 3DVLE represents a time- consuming stage in the overall pipeline production. One reason is that a range of technologies and tools are typically used. In (Cudworth 2014) a 3-week period is indicated for experienced users to perform the 3D modeling of a virtual space. In our case, a 3-month work was needed for designing a working model of a 3D virtual campus (see Figure 1 and Figure 2 for final results).

opencc-by-4.0Dec 2017View details →

Prototyping 3D Virtual Learning Environments with X3D-based Content and Visualization Tools-Figure 1. The model of the 3D virtual campus - an outdoor view (3D modeling by Marius Hodea)

Prototyping 3D Virtual Learning Environments with X3D-based Content and Visualization Tools-Figure 9. Room model generated with Autodesk 123D Catch - the 2D model

Prototyping 3D Virtual Learning Environments with X3D-based Content and Visualization Tools-Figure 8. The 3D model of the faculty building in a HTML page

The pipeline processing was the following: a) the 3D model from Sketchup was saved as a Collada file; b) this file has been imported in MeshLab (MESHLAB 2017) and converted to VRML97 format (wrl); c) aopt utility was used to convert wrl files to X3D and HTML5 files. The model was visualized in the OpenSim virtual world setting using an external browser (Figure 8).

Prototyping 3D Virtual Learning Environments with X3D-based Content and Visualization Tools-Figure 5. Iteration of the 3D modeling of the virtual faculty building (3D modeling by Marius Hodea)

For our research, several iterations and methods were employed for the 3D design of an online campus.  Different virtual models of a faculty building (see Figures 4,5) were designed and finally a virtual model of a 3D campus comprising a simplified 3-story faculty building (Figure 6) was created. The objective was the optimization of the 3D model and the demonstration of the desired functionalities. For these purposes two 3D modeling and post-processing software were used, i.e. 3DSMax and Trimble Sketchup. The model of the building resulted in 5962 vertices and 4528 faces. The textures and illumination were applied using OpenSim’s in-world tools. Furniture objects (tables, chair, computer monitors) were taken from the Google 3D Warehouse, distributed and shared under Trimble General Model License.

Prototyping 3D Virtual Learning Environments with X3D-based Content and Visualization Tools-Figure 11. Online accessible repository of digital data on cultural heritage with X3D models (STARC Web Repository, 2017, © Copyright 2017, STARC, Cyprus Institute. Used with permission)

Prototyping can also include the development of toolkits for automatic content generation simulator, but in the case of an architectural environment, the components are too complex to be automatically generated. Furniture elements or the learning artifacts (i.e. content created by learners) can be converted to be viewed in X3D compatible browsers or included in online galleries (Figure 11). After functional and 3D content prototyping, certain components of the virtual campus can be easily modified and adapted as needed.

Prototyping 3D Virtual Learning Environments with X3D-based Content and Visualization Tools-Figure 6. The first implementation of the virtual campus (OpenSim import of the 3D model)

Prototyping 3D Virtual Learning Environments with X3D-based Content and Visualization Tools-Figure 7. The HTML source (partial) code, integrating the X3D model

To integrate the model into a web page, a model conversion to X3D format and an X3DOM output under the form of an HTML5 encoded webpage (Figure 7) were needed. Instant Reality distribution provides a command line transcoding tool, named Avalon Optimizer (aopt), that was used to convert a VRML format (wrl extension) of the model to X3D.