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.
129
datasets available to search
ShareScore release 0.9.0
Dataset results
129 results for “Deep neural network”
Automated detection, segmentation and classification of pericardial effusions on chest CT using a deep convolutional neural network
<p>Trainingsdata for chest CT pericard effusion and the finish trained nnU-Net model. </p>
Explaining reaction coordinates of alanine dipeptide isomerization obtained from deep neural networks using Explainable Artificial Intelligence (XAI)
<p>This repository includes the input and output dataset, and python scripts used in the article, "Explaining reaction coordinates of alanine dipeptide isomerization obtained from deep neural networks using Explainable Artificial Intelligence (XAI)," of J. Chem. Phys. 156, 154108 (2022) [DOI: <a href="http://doi.org/10.1063/5.0087310">10.1063/5.0087310</a>] The repository also includes source data of figures in the article.</p>
Research Artifact for paper "Decompiling x86 Deep Neural Network Executables"
<p>Research Artifact for USENIX Security 2023 paper "Decompiling x86 Deep Neural Network Executables"</p>
Predicting Code Comprehension: A Novel Approach to Align Human Gaze with Code Using Deep Neural Networks
<p><strong>Checkout our Github-Repo for more information, issues, and pull requests: </strong></p> <p><a href="https://github.com/Taremeh/predicting-code-comprehension-eye-tracking/">https://github.com/Taremeh/predicting-code-comprehension-eye-tracking/</a></p> <p> </p> <p>Dataset and Replication Package for our paper "Predicting Code Comprehension: A Novel Approach to Align Human Gaze with Code Using Deep Neural Networks"</p>
DeepFruits: A Fruit Detection System Using Deep Neural Networks
<p>This is the dataset associated with MDPI Sensors paper entitled "DeepFruits: A Fruit Detection System Using Deep Neural Networks".</p> <p>This paper presents a novel approach to fruit detection using deep convolutional neural networks. The aim is to build an accurate, fast and reliable fruit detection system, which is a vital element of an autonomous agricultural robotic platform; it is a key element for fruit yield estimation and automated harvesting. Recent work in deep neural networks has led to the development of a state-of-the-art object detector termed Faster Region-based CNN (Faster R-CNN). We adapt this model, through transfer learning, for the task of fruit detection using imagery obtained from two modalities: colour (RGB) and Near-Infrared (NIR). Early and late fusion methods are explored for combining the multi-modal (RGB and NIR) information. This leads to a novel multi-modal Faster R-CNN model, which achieves state-of-the-art results compared to prior work with the F1 score, which takes into account both precision and recall performances improving from 0.807 to 0.838 for the detection of sweet pepper. In addition to improved accuracy, this approach is also much quicker to deploy for new fruits, as it requires bounding box annotation rather than pixel-level annotation (annotating bounding boxes is approximately an order of magnitude quicker to perform). The model is retrained to perform the detection of seven fruits, with the entire process taking four hours to annotate and train the new model per fruit.</p>
Global Prediction Of Total Organic Carbon In Marine Sediments Using Deep Neural Networks (nn-toc) v2
<p>This is the second version that was uploaded to make the code available for the paper submission <strong><span>NN-TOC v1: global prediction of total organic carbon in marine sediments using deep neural networks</span></strong> to the Geoscientific Model Development journal. Here we create a deep neural network based approach for the geospatial predicition of total organic carbon percentages in marine sediments.</p> <p><span>The data folder contains "raw" features and labels, "interim" data for preprocessed features and labels and "output"s produced from the model. While the preprocessed folder contain all the other files that can be produced by running the code. The features are in .nc or .grd file format. The other files are in .xyz or .csv file format.</span></p> <p> </p>
Dataset used in "Estimating Dispersion Coefficient in Flow Through Heterogeneous Porous Media by a Deep Convolutional Neural Network" by Kamrava et al. in Geophysical Research Letters.
<p>Morphology of Heterogeneous Porous Media</p>
Trained deep neural networks for MSI/dMMR detection in colorectal cancer histology
<p>These are trained neural network models in PyTorch format to process tessellated images of colorectal cancer histology samples. The input is expected to be 224x224 px RGB image tiles normalized with the Macenko method. The output is a probability of the image tile for being MSI/dMMR or MSS/pMMR. </p> <p>The models have been trained on eight cohorts but not on the validation cohort. The validation cohorts are:</p> <p>Ex_0 : DACHS</p> <p>Ex_1 : DUSSEL</p> <p>Ex_2 : MECC</p> <p>Ex_3 : QUASAR</p> <p>Ex_4 : RAINBOW</p> <p>Ex_5 : TCGA</p> <p>Ex_6 : UMM</p> <p>Ex_7 : YORKSHIRE</p> <p>Ex_8 : MUNICH</p> <p>The models can be loaded in Python with </p> <p>>>> model = torch.load(path, map_location=torch.device('cpu'))</p> <p>Further details are given in the manuscript.</p>
Screening for Chagas disease using a deep neural network
<p>Pre-trained models accompanying the source code at https://github.com/carji475/ecg-chagas</p>
Automated detection, segmentation and classification of pericardial effusions on chest CT using a deep convolutional neural network
<p>Test data set</p>
DeepC: Predicting chromatin interactions using megabase scaled deep neural networks and transfer learning (NG Capture-C)
GEO Series GSE137435. Homo sapiens. 4 samples. Type: Other.
DeepC: Predicting chromatin interactions using megabase scaled deep neural networks and transfer learning (Tiled-C)
GEO Series GSE137436. Homo sapiens. 2 samples. Type: Other.
Visual perception of liquids: insights from deep neural networks
<p>Datasets and analysis code of the following publication:</p> <p>Van Assen, J.J.R., Nishida, S. & Fleming, R. W. (2020). Visual perception of liquids: insights from deep neural networks. <em>PLOS Computational Biology. </em>DOI: 10.1371/journal.pcbi.1008018</p> <p>For any questions please contact the first author at mail [at] janjaap [dot] info</p> <p><strong>Contents:</strong></p> <p>1. DataAnalysis<br> - Jupyter Notebook to run the full analysis in R<br> - For installation details see: https://irkernel.github.io/requirements/</p> <p>2. FullStimulusSet<br> - 2 million liquid images with 16 viscosities, 10 scenes, 625 variations, and 20 frames<br> - Matlab script that merges the images horizontally for network input</p> <p>3. NeuralActivations<br> - Matlab files containing the neural activations if you cannot read out the networks</p> <p>4. TrainedNetworks<br> - 100 Trained networks referred to in the paper using Matlab and the Deep Learning Toolbox<br> - One custom layer file “switchLayerAdvanced.m”</p> <p>5. ValidationSet<br> - 800 experimental stimuli that were used for validation 16 viscosities, 10 scenes, 5 variations (1,6,11,16,21)<br> - Matlab script that merges the images horizontally for network input</p>
Interpretable Deep Convolutional Neural Network-based Surrogates for Complex Urban Hydrodynamic Modeling using Random Chaotic Rainfall
<p>This data is part of the article "Interpretable Deep Convolutional Neural Network-based Surrogates for Complex Urban Hydrodynamic Modeling using. Random Chaotic Rainfall "data for model training and validation, as well as a dynamic runoff range generated by DHMUrban</p>
[LS2N_IPI_DisFER] Comparing the Robustness of Humans and Deep Neural Networks on Facial Expression Recognition
<h2>DisFER</h2> <p>Distorted-FER (DisFER), a new facial expression recognition (FER) dataset composed of a wide number of distorted images of faces.</p> <p> </p> <h2>Materials and Methods</h2> <h4>Dataset</h4> <div>The source images used in our experiment come from the Facial Expression Recognition 2013 (FER-2013) dataset [1]. This dataset was firstly introduced in 2013 at the International Conference on Machine Learning, and has been used in a large number of research works since then, as it encompasses naturalistic conditions and challenges. This dataset consists of 35,887 images of faces in 48 × 48 format, collected thanks to a Google search. Human accuracy on FER-2013 was estimated by its authors around 65.5% [1].</div> <div>To build the Distorted-FER (DisFER) dataset, we randomly selected, from FER-2013, twelve images per basic emotion, as defined by Ekman [2] (i.e., anger, disgust, fear, happiness, neutral, sadness, and surprise). This yields a total of 84 source images. Each original stimulus was then distorted using three different types of distortions, i.e., Gaussian blur (GB), Gaussian noise (GN), and salt-and-pepper noise (SP). Each distortion was applied at distinct levels: three standard deviation values were tested for GB, i.e., 0.8, 1.1, and 1.4; similarly for GN with standard deviation values equal to 10, 20, and 30; while probability levels of 0.02, 0.04, and 0.06 were chosen for SP; corresponding to low, medium, and high distortions, respectively. </div> <div> </div> <div> <h4>Crowdsourcing Experiment</h4> <div>In order to collect as many votes as possible on our dataset, and because rating 840 images is time-consuming and can be extremely tiring for a single participant, we decided to set up a crowdsourcing experiment. Such experiments indeed allow the conduct of large-scale subjective tests with reduced costs and efforts.</div> <div>The DisFER dataset was therefore split into twenty-one playlists of forty images each, with a view to keep the tests as fast as possible—as crowdsourcing experiments should not last more than ten minutes or so. Playlists were carefully designed to contain the same numbers of images of a given configuration (i.e., emotion, distortion types, and distortion levels). Among a playlist, images were randomly displayed to participants.</div> <div>Each participant was asked to choose which emotion (i.e., anger, disgust, fear, happiness, neutral, sadness, or surprise) they recognized in the displayed image. No time constraint was imposed on participants to fulfill the task.</div> <div> </div> <div>A total of 1051 participants (including 50% of females) were recruited using the Prolific platform [3]. Prolific takes into consideration researchers’ needs by maintaining a subject recruitment process that is similar to that of a laboratory experiment. Indeed, participants are fully informed that they are being recruited for a research study. Consequently, this platform allows researchers to eliminate ethical concerns, and it further improves the reliability of collected data.</div> <div> </div> Participants were aged between 19 and 75 years old (with a mean of 30±8.53 -- note that three participants did not wish to respond). Twenty playlists out of twenty-one were entirely watched and rated by fifty distinct participants, whereas one playlist was watched and evaluated by fifty-one participants.</div> <div> </div> <div> <h2>References</h2> </div> <div>[1] Goodfellow, I.J.; Erhan, D.; Carrier, P.L.; Courville, A. Challenges in Representation Learning: A Report on Three Machine Learning Contests. In Proceedings of the Neural Information Processing, Daegu, South Korea, 3–7 November 2013; Springer: Berlin/Heidelberg, Germany, 2013; pp. 117–124</div> <div>[2] Ekman, P. An argument for basic emotions. Cogn. Emot. <strong>1992</strong>, 6, 169–200</div> <div>[3] https://www.prolific.com/</div> <div> </div>
Mutate and Observe: Utilizing Deep Neural Networks to Investigate the Impact of Mutations on Translation Initiation
<p>Datasets used in the paper.</p>
Med-ReLU: A Hybrid Activation Function Tailored for Deep Artificial Neural Networks in Medical Image Segmentation without Parameter Tuning
<p>Background: Deep learning (DL) is derived from the domain of Artificial Neural Network (ANN). It makes one of the most important elements of deep learning algorithms. Deep learning segmentation models are based on layer-by-layer convolution learning attribute representation directed by forward and backward propagation. Throughout the process vital role is played by appropriately chosen activation function (AF) in order to guarantee the robustness of the model learning. However, the existing activation functions are either ineffective in addressing the vanishing gradient problem or get burdened with multiple parameters that need to be manually tuned. Moreover, the current research on activation function design mainly focuses on classification tasks using natural images from the MNIST, CIFAR-10 and CIFAR-100 datasets. Therefore,Med-ReLU as a novel activation function for medical image segmentation, is proposed. The proposed activation function avoids deep learning models from the attacks of dead neurons or from the vanishing gradient problems. Method: Med-ReLU is a hybrid activation function that combines the property of two activation functions of ReLU and Softsign. For positive inputs, Med-ReLU utilizes the linear property just like ReLU to produce an output without vanishing gradient. The negative inputs converge in polynomial ways towards their asymptotes as property of the softsign AF that ensures robust training processing without the problem of dead neurons that rarely activate across the entire training dataset. Results: The training performance and segmentation accuracy of Med-ReLU have been investigated. The proposed function has demonstrated stable training and does not suffer from over-fitting. Hence, Med-ReLU has consistently outperformed the existing state-of-art activation functions in medical image segmentation tasks. Conclusion: Med-ReLU has been designed as a parameter-free activation function for DL image segmentation tasks. This activation function is easy-to-implement on complex and deep learning models. The utility of this research lies in affirming the impact of Med-ReLU on different Artificial Neural Network architectures and for various kinds of anomaly addressing tasks.</p>
Positron Emission Tomography (PET) Images Using Deep Neural Networks
ClinicalTrials.gov study NCT04140565. IPD Sharing: NO. Countries: 1. Publications: 0.
Deep Neural Network Approaches for Closed-Loop Deep Brain Stimulation
ClinicalTrials.gov study NCT04277689. IPD Sharing: NO. Countries: 1. Publications: 0.
Research on the Risk Warning Model and Prevention Strategies for Acute Kidney Injury Associated With Cyclosporine Based on Explainable Deep Neural Networks and Therapeutic Drug Monitoring
ClinicalTrials.gov study NCT06596811. IPD Sharing: Not stated. Countries: 1. Publications: 0.
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.