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.
282
datasets available to search
ShareScore release 0.7.1
Dataset results
282 results for “Image segmentation”
Pre-trained models for segmentation and tracking of Coronal Bright Fronts from SDO AIA Base Difference images
<p>Here we present pretrained U-NET-based models followed by SDO AIA Base Difference(BD) validation set after intensity tresholding [-50;150] with predicted feature masks samples. <br>We provide a command-line Python utility for image segmentation using our CNNs designed to process images of solar eruptive phenomena. The https://gitlab.com/iahelio/helios_cnn repository includes regularly updated and newly published models. </p> <p>First model we present is designed to predict the likelihood of each pixel belonging to a certain class or feature in the solar image. A probabilistic output allows for a more nuanced interpretation of ambiguous region. The output can be converted into binary masks through thresholding. The range of values also gives insights into the model's confidence</p> <p>We also present sample segmentation results and the second model designed to produce binary masks.</p>
Ultrafast imaging recordings from the axon initial segment of neocortical layer-5 pyramidal neurons.
<p>This dataset contains imaging and whole-cell electrophysiological recordings from neocortical layer-5 pyramidal neuron from brain slices of the mouse.</p> <p>Electrophysiological recordings (at 20 kHz) are from the soma. Imaging data (10 kHz) are from lines along the axon initial segment (distal>proximal) with 500 nm pixel resolution. These correspond to:</p> <ul> <li>Sodium imaging (Figures 1 and S6).</li> <li>Voltage imaging (Figures 2,4,5,S4,S7)</li> <li>Calcium imaging (Figures 3,S3,S8).</li> </ul> <p>This dataset is used in the paper available online:</p> <p>Filipis L, Blömer LA, Montnach J, De Waard M, Canepari M. Nav1.2 and BK channels interaction shapes the action potential in the axon initial segment. bioRxiv, 2022. doi: 10.1101/2022.04.12.488116.</p>
S1S2-Water: A global dataset for semantic segmentation of water bodies from Sentinel-1 and Sentinel-2 satellite images
<p>The S1S2-Water dataset is a global reference dataset for training, validation and testing of convolutional neural networks for semantic segmentation of surface water bodies in publicly available Sentinel-1 and Sentinel-2 satellite images. The dataset consists of 65 triplets of Sentinel-1 and Sentinel-2 images with quality checked binary water mask. Samples are drawn globally on the basis of the Sentinel-2 tile-grid (100 x 100 km) under consideration of pre-dominant landcover and availability of water bodies. Each sample is complemented with metadata and Digital Elevation Model (DEM) raster from the Copernicus DEM.</p><p>This work was supported by the German Federal Ministry of Education and Research (BMBF) through the project "Künstliche Intelligenz zur Analyse von Erdbeobachtungs- und Internetdaten zur Entscheidungsunterstützung im Katastrophenfall" (AIFER) under Grant 13N15525, and by the Helmholtz Artificial Intelligence Cooperation Unit through the project "AI for Near Real Time Satellite-based Flood Response" (AI4FLOOD) under Grant ZT-IPF-5-39. </p>
Image segmentations produced by BAMF under the AIMI Annotations initiative
<p>The Imaging Data Commons (IDC)(<a href="https://imaging.datacommons.cancer.gov/">https://imaging.datacommons.cancer.gov/</a>) [1] connects researchers with publicly available cancer imaging data, often linked with other types of cancer data. Many of the collections have limited annotations due to the expense and effort required to create these manually. The increased capabilities of AI analysis of radiology images provide an opportunity to augment existing IDC collections with new annotation data. To further this goal, we trained several nnUNet [2] based models for a variety of radiology segmentation tasks from public datasets and used them to generate segmentations for IDC collections.</p> <p>To validate the model's performance, roughly 10% of the AI predictions were assigned to a validation set. For this set, a board-certified radiologist graded the quality of AI predictions on a Likert scale. If they did not 'strongly agree' with the AI output, the reviewer corrected the segmentation. </p> <p>This record provides the AI segmentations, Manually corrected segmentations, and Manual scores for the inspected IDC Collection images.</p> <p><em>Only 10% of the AI-derived annotations provided in this dataset are verified by expert radiologists . More details, on model training and annotations are provided within the associated manuscript to ensure transparency and reproducibility.</em></p> <p> </p> <p>This work was done in two stages. Versions 1.x of this record were from the first stage. Versions 2.x added additional records. In the Version 1.x collections, a medical student (non-expert) reviewed all the AI predictions and rated them on a 5-point Likert Scale, for any AI predictions in the validation set that they did not 'strongly agree' with, the non-expert provided corrected segmentations. This non-expert was not utilized for the Version 2.x additional records.</p> <p> </p> <h3>Likert Score Definition:</h3> <p>Guidelines for reviewers to grade the quality of AI segmentations.</p> <ul> <li>5 Strongly Agree - Use-as-is (i.e., clinically acceptable, and could be used for treatment without change)</li> <li>4 Agree - Minor edits that are not necessary. Stylistic differences, but not clinically important. The current segmentation is acceptable</li> <li>3 Neither agree nor disagree - Minor edits that are necessary. Minor edits are those that the review judges can be made in less time than starting from scratch or are expected to have minimal effect on treatment outcome</li> <li>2 Disagree - Major edits. This category indicates that the necessary edit is required to ensure correctness, and sufficiently significant that user would prefer to start from the scratch</li> <li>1 Strongly disagree - Unusable. This category indicates that the quality of the automatic annotations is so bad that they are unusable.</li> </ul> <p> </p> <h3>Zip File Folder Structure</h3> <p>Each zip file in the collection correlates to a specific segmentation task. The common folder structure is</p> <ul> <li><em>ai-segmentations-dcm </em>This directory contains the AI model predictions in DICOM-SEG format for all analyzed IDC collection files</li> <li><em>qa-segmentations-dcm </em>This directory contains manual corrected segmentation files, based on the AI prediction, in DICOM-SEG format. Only a fraction, ~10%, of the AI predictions were corrected. Corrections were performed by radiologist (rad*) and non-experts (ne*)</li> <li><em>qa-results.csv</em> CSV file linking the study/series UIDs with the ai segmentation file, radiologist corrected segmentation file, radiologist ratings of AI performance.</li> </ul> <p> </p> <h3><strong><em>qa-results.csv Columns</em></strong></h3> <p>The qa-results.csv file contains metadata about the segmentations, their related IDC case image, as well as the Likert ratings and comments by the reviewers.</p> <div> <table> <tbody> <tr> <td> <p><strong>Column</strong></p> </td> <td> <p><strong>Description</strong></p> </td> </tr> <tr> <td> <p><em>Collection</em></p> </td> <td> <p>The name of the IDC collection for this case</p> </td> </tr> <tr> <td> <p><em>PatientID</em></p> </td> <td> <p>PatientID in DICOM metadata of scan. Also called Case ID in the IDC</p> </td> </tr> <tr> <td> <p><em>StudyInstanceUID</em></p> </td> <td> <p>StudyInstanceUID in the DICOM metadata of the scan</p> </td> </tr> <tr> <td> <p><em>SeriesInstanceUID</em></p> </td> <td> <p>SeriesInstanceUID in the DICOM metadata of the scan</p> </td> </tr> <tr> <td> <p><em>Validation</em></p> </td> <td> <p>true/false if this scan was manually reviewed</p> </td> </tr> <tr> <td> <p><em>Reviewer</em></p> </td> <td> <p>Coded ID of the reviewer. Radiologist IDs start with ‘rad’ non-expect IDs start with ‘ne’</p> </td> </tr> <tr> <td> <p><em>AimiProjectYear</em></p> </td> <td> <p>2023 or 2024, This work was split over two years. The main methodology difference between the two is that in 2023, a non-expert also reviewed the AI output, but a non-expert was not utilized in 2024.</p> </td> </tr> <tr> <td> <p><em>AISegmentation</em></p> </td> <td> <p>The filename of the AI prediction file in DICOM-seg format. This file is in the ai-segmentations-dcm folder.</p> </td> </tr> <tr> <td> <p><em>CorrectedSegmentation</em></p> </td> <td> <p>The filename of the reviewer-corrected prediction file in DICOM-seg format. This file is in the qa-segmentations-dcm folder. If the reviewer strongly agreed with the AI for all segments, they did not provide any correction file.</p> </td> </tr> <tr> <td> <p><em>Was the AI predicted ROIs accurate?</em></p> </td> <td> <p>This column appears one for each segment in the task for images from AimiProjectYear 2023. The reviewer rates segmentation quality on a Likert scale. In tasks that have multiple labels in the output, there is only one rating to cover them all.</p> </td> </tr> <tr> <td> <p><em>Was the AI predicted {SEGMENT_NAME} label accurate?</em></p> <p><em><strong> </strong></em></p> </td> <td> <p>This column appears one for each segment in the task for images from AimiProjectYear 2024. The reviewer rates each segment for its quality on a Likert scale.</p> </td> </tr> <tr> <td> <p><em>Do you have any comments about the AI predicted ROIs?</em></p> <p><em><strong> </strong></em></p> </td> <td> <p>Open ended question for the reviewer</p> </td> </tr> <tr> <td> <p><em>Do you have any comments about the findings from the study scans?</em></p> </td> <td> <p>Open ended question for the reviewer</p> </td> </tr> </tbody> </table> </div> <p> </p> <h3>File Overview</h3> <h4>brain-mr.zip</h4> <ul> <li>Segment Description: brain tumor regions: necrosis, edema, enhancing</li> <li>IDC Collection: <a href="https://www.cancerimagingarchive.net/collection/upenn-gbm/">UPENN-GBM</a></li> <li>Links: <a href="../records/11582627">model weights</a>, <a href="https://github.com/bamf-health/aimi-brain-mr">github</a></li> </ul> <h4>breast-fdg-pet-ct.zip</h4> <ul> <li>Segment Description: FDG-avid lesions in breast from FDG PET/CT scans QIN-Breast</li> <li>IDC Collection: <a href="https://www.cancerimagingarchive.net/collection/qin-breast/">QIN-Breast</a></li> <li>Links: <a href="https://doi.org/10.5281/zenodo.8290054">model weights, </a><a href="https://github.com/bamf-health/aimi-breast-pet-ct">github</a></li> </ul> <h4>breast-mr.zip</h4> <ul> <li>Segment Description: Breast, Fibroglandular tissue, structural tumor</li> <li>IDC Collection: <a href="https://www.cancerimagingarchive.net/collection/duke-breast-cancer-mri/">duke-breast-cancer-mri</a></li> <li>Links: <a href="../records/11998679">model weights</a>, <a href="https://github.com/bamf-health/aimi-breast-mr">github</a></li> </ul> <h4>kidney-ct.zip</h4> <ul> <li>Segment Description: Kidney, Tumor, and Cysts from contrast enhanced CT scans</li> <li>IDS Collection: <a href="https://www.cancerimagingarchive.net/collection/tcga-kirc/">TCGA-KIRC,</a> <a href="https://www.cancerimagingarchive.net/collection/tcga-kirp/">TCGA-KIRP</a>, <a href="https://www.cancerimagingarchive.net/collection/tcga-kich/">TCGA-KICH</a>, <a href="https://www.cancerimagingarchive.net/collection/cptac-ccrcc/">CPTAC-CCRCC</a></li> <li>Links: <a href="https://doi.org/10.5281/zenodo.8277845">model weights, </a><a href="https://github.com/bamf-health/aimi-kidney-ct">github</a></li> </ul> <h4>liver-ct.zip</h4> <ul> <li>Segment Description: Liver from CT scans</li> <li>IDC Collection: <a href="https://www.cancerimagingarchive.net/collection/TCGA-LIHC/">TCGA-LIHC</a></li> <li>Links: <a href="https://doi.org/10.5281/zenodo.8270230">model weights, </a><a href="https://github.com/bamf-health/aimi-liver-ct">github</a></li> </ul> <h4>liver2-ct.zip</h4> <ul> <li>Segment Description: Liver and Lesions from CT scans</li> <li>IDC Collection: <a href="https://www.cancerimagingarchive.net/collection/hcc-tace-seg/">HCC-TACE-SEG</a>, <a href="https://www.cancerimagingarchive.net/collection/colorectal-liver-metastases/">COLORECTAL-LIVER-METASTASES</a></li> <li>Links: <a href="../records/11582728">model weights</a>, <a href="https://github.com/bamf-health/aimi-liver-tumor-ct">github</a></li> </ul> <h4>liver-mr.zip</h4> <ul> <li>Segment Description: Liver from T1 MRI scans</li> <li>IDC Collection: <a href="https://www.cancerimagingarchive.net/collection/TCGA-LIHC/">TCGA-LIHC</a></li> <li>Links: <a href="https://doi.org/10.5281/zenodo.8290123">model weights, </a><a href="https://github.com/bamf-health/aimi-liver-mr">github</a></li> </ul> <h4>lung-ct.zip</h4> <ul> <li>Segment Description: Lung and Nodules (3mm-30mm) from CT scans</li> <li>IDC Collections:<br> <ul> <li><a href="https://www.cancerimagingarchive.net/collection/anti-pd-1_lung/">Anti-PD-1-Lung</a></li> <li><a href="https://www.cancerimagingarchive.net/collection/lung-pet-ct-dx/">LUNG-PET-CT-Dx</a></li> <li><a href="https://www.cancerimagingarchive.net/collection/nsclc-radiogenomics/">NSCLC Radiogenomics</a></li> <li><a href="https://www.cancerimagingarchive.net/collection/rider-lung-pet-ct/">RIDER Lung PET-CT</a></li> <li><a href="https://www.cancerimagingarchive.net/collection/TCGA-LUAD/">TCGA-LUAD</a></li> <li><a href="https://www.cancerimagingarchive.net/collection/TCGA-LUSC/">TCGA-LUSC</a></li> </ul> </li> <li>Links: <a href="https://doi.org/10.5281/zenodo.8290146">model weights 1, </a><a href="../record/8290169">model weights 2, </a><a href="https://github.com/bamf-health/aimi-lung-ct">github</a></li> </ul> <h4>lung2-ct.zip</h4> <ul> <li>Improved model version</li> <li>Segment Description: Lung and Nodules (3mm-30mm) from CT scans</li> <li>IDC Collections:<br> <ul> <li><a href="https://www.cancerimagingarchive.net/collection/QIN-LUNG-CT">QIN-LUNG-CT</a>, <a href="https://www.cancerimagingarchive.net/collection/spie-aapm-lung-ct-challenge/">SPIE-AAPM Lung CT Challenge</a></li> </ul> </li> <li>Links: <a href="../records/11582738">model weights</a>, <a href="https://github.com/bamf-health/aimi-lung2-ct">github</a></li> </ul> <h4>lung-fdg-pet-ct.zip</h4> <ul> <li>Segment Description: Lungs and FDG-avid lesions in the lung from FDG PET/CT scans</li> <li>IDC Collections: <ul> <li><a href="https://www.cancerimagingarchive.net/collection/acrin-nsclc-fdg-pet/">ACRIN-NSCLC-FDG-PET</a></li> <li><a href="https://www.cancerimagingarchive.net/collection/anti-pd-1_lung/">Anti-PD-1-Lung</a></li> <li><a href="https://www.cancerimagingarchive.net/collection/lung-pet-ct-dx/">LUNG-PET-CT-Dx</a></li> <li><a href="https://www.cancerimagingarchive.net/collection/nsclc-radiogenomics/">NSCLC Radiogenomics</a></li> <li><a href="https://www.cancerimagingarchive.net/collection/rider-lung-pet-ct/">RIDER Lung PET-CT</a></li> <li><a href="https://www.cancerimagingarchive.net/collection/TCGA-LUAD/">TCGA-LUAD</a></li> <li><a href="https://www.cancerimagingarchive.net/collection/TCGA-LUSC/">TCGA-LUSC</a></li> </ul> </li> <li>Links: <a href="https://doi.org/10.5281/zenodo.8290054">model weights, </a><a href="https://github.com/bamf-health/aimi-lung-pet-ct">github</a></li> </ul> <h4>prostate-mr.zip</h4> <ul> <li>Segment Description: Prostate from T2 MRI scans</li> <li>IDC Collection: <a href="https://www.cancerimagingarchive.net/collection/ProstateX/">ProstateX,</a> <a href="https://www.cancerimagingarchive.net/collection/prostate-mri-us-biopsy/">Prostate-MRI-US-Biopsy</a></li> <li>Links: <a href="https://doi.org/10.5281/zenodo.8290092">model weights, </a><a href="https://github.com/bamf-health/aimi-prostate-mr">github</a></li> </ul> <p> </p> <p><strong>Changelog</strong></p> <ul> <li>2.0.2 - Fix the brain-mr segmentations to be transformed correctly</li> <li>2.0.1 - added AIMI 2024 radiologist comments to qa-results.csv</li> <li>2.0.0 - added AIMI 2024 segmentations</li> <li>1.X - AIMI 2023 segmentations and reviewer scores</li> </ul>
Doodleverse/Segmentation Zoo Res-UNet model for NOAA ERI/4-class segmentation of RGB 512x512 images
<p>This Residual-UNet model is trained on 1,179 pairs of human-generated segmentation labels and images from Emergency Response Imagery (ERI) collected by US National Oceanic and Atmospheric Administration (NOAA) after Hurricane Barry, Delta, Dorian, Florence, Ida, Laura, Michael, Sally, Zeta, and Tropical Storm Gordon. The dataset is available here: https://doi.org/10.5281/zenodo.7268082</p> <p>Models have been created using Segmentation Gym:</p> <p>Code - https://github.com/Doodleverse/segmentation_gym</p> <p>Paper - https://doi.org/10.1029/2022EA002332</p> <p> </p> <p>The model takes input images that are 512 x 512 x 3 pixels, and the output is 512 x 512 x 4, corresponding to 4 classes:</p> <ol> <li>water</li> <li>bare sediment</li> <li>vegetation</li> <li>development (roads, buildings, power lines, parking lots, etc.)</li> </ol> <p> </p> <p>Included here are 6 files with the same root name:</p> <ol> <li> '.json' config file: this is the file that was used by Segmentation Gym to create the weights file. It contains instructions for how to make the model and the data it used, as well as instructions for how to use the model for prediction.</li> <li>'.h5' weights file: this is the file that was created by the Segmentation Gym function `train_model.py`. It contains the trained model's parameter weights. It can called by the Segmentation Gym function `seg_images_in_folder.py`.</li> <li> '_model_history.npz' model training history file: this numpy archive file contains numpy arrays describing the training and validation losses and metrics. It is created by the Segmentation Gym function `train_model.py`</li> <li> '.png' model training loss and mean IoU plot: this png file contains plots of training and validation losses and mean IoU scores during model training. A subset of data inside the .npz file. It is created by the Segmentation Gym function `train_model.py`</li> <li>'.zip' of the model in the Tensorflow ‘saved model’ format. It is created by the Segmentation Gym function `utils/gen_saved_model.py`</li> <li>'_modelcard.json' model card file: this is a json file containing fields that collectively describe the model origins, training choices, and dataset that the model is based upon. There is some redundancy between this file and the `config` file (described above) that contains the instructions for the model training and implementation. The model card file is not used by the program but is important metadata so it is important to keep with the other files that collectively make the model and is such is considered part of the model</li> </ol> <p>Additionally, BEST_MODEL.txt contains the name of the model with the best validation loss and mean IoU</p>
Pre-training with simulated ultrasound images for breast mass segmentation and classification - dataset
<p>Dataset assosiated with the MICCAI Workshop on Data Engineering in Medical Imaging paper: "Pre-training with Simulated Ultrasound Images for Breast Mass Segmentation and Classification"</p>
Artefact segmentation in digital pathology whole-slide images
<p>Dataset with examples of Artefacts in Digital Pathology.</p> <p>The dataset contains 22 Whole-Slide Images, with H&E or IHC staining, showing various types and levels of defect to the slides. Annotations were made by a biomedical engineer based on examples given by an expert.</p> <p>The dataset is split in different folders:</p> <ul> <li>train <ul> <li>18 whole-slide images (extracted at 1.25x & 2.5x magnification)</li> <li>All from the same Block (colorectal cancer tissue)</li> <li>1/2 with H&E & 1/2 with anti-pan-cytokeratin IHC staining.</li> </ul> </li> <li>validation <ul> <li>3 whole-slide images (1.25x + 2.5x mag)</li> <li>2 from the same Block as the training set (1 IHC, 1 H&E)</li> <li>1 from another Block (IHC anti-pan-cytokerating, gastroesophageal junction lesion)</li> </ul> </li> <li>validation_tiles <ul> <li>patches of varying sizes taken from the 3 validation whole-slide images @1.25x magnification.</li> <li>7 patches from each slide.</li> </ul> </li> <li>test <ul> <li>1 whole-slide image (1.25x + 2.5x mag)</li> <li>From another block: IHC staining (anti-NR2F2), mouth cancer</li> </ul> </li> </ul> <p>For the train, validation and test whole-slide images, each slide has:<br> - The RGB images @1.25x & 2.5x mag<br> - The corresponding background/tissue masks<br> - The corresponding annotation masks containing examples of artefacts (note that a majority of artefacts are not annotated. In total, 918 artefacts are in the train set)</p> <p>For the validation tiles, the following table gives the "patch-level" supervision:</p> <p>tile# Artefact(s)<br> 00 None/Few<br> 01 Tear&Fold<br> 02 Ink<br> 03 None/Few<br> 04 None/Few<br> 05 Tear&Fold<br> 06 Tear&Fold + Blur<br> 07 Knife damage<br> 08 Knife damage<br> 09 Ink<br> 10 None/Few<br> 11 Tear&Fold<br> 12 Tear&Fold<br> 13 None/Few<br> 14 None/Few<br> 15 Knife damage<br> 16 Tear&Fold<br> 17 None/Few<br> 18 None/Few<br> 19 Blur<br> 20 Knife damage</p>
Cu dataset – A copper ore labeled images dataset for segmentation training and testing
<p>This dataset is composed of 121 pairs of correlated images. Each pair contains one image of a copper ore sample acquired through reflected light microscopy (RGB, 24-bit), and the corresponding binary reference image (8-bit), in which the pixels are labeled as belonging to one of two classes: ore (0) or embedding resin (255).</p> <p>The sample came from a copper ore from Yauri Cusco (Peru) with a complex mineralogy, mainly composed of sulfides, oxides, silicates, and native copper. It was classified by size. The fraction +74-100 μm was cold mounted with epoxy resin and subsequently ground and polished.</p> <p>Correlative microscopy was employed for image acquisition. Thus, 121 fields were imaged on a reflected light microscope with a 20× (NA 0.40) objective lens and on a scanning electron microscope (SEM). In sequence, they were registered, resulting in images of 1017×753 pixels with a resolution of 0.53 µm/pixel. As matter of fact, some images (the images No. 2, 3, 24, 25, 46, 47, 69, 91, and 113) have slightly smaller sizes because they were cropped during the registration procedure to correct co-localization errors of the order of a few pixels. Finally, the images from SEM were thresholded to generate the reference images.</p> <p>Further description of this sample and its imaging procedure can be found in the work by Gomes and Paciornik (2012).</p> <p>This dataset was created for developing and testing deep learning models on semantic segmentation tasks. The paper of Filippo et al. (2021) presented a variant of the DeepLabv3+ model (Chen et al., 2018) that reached mean values of 90.56% and 92.12% for overall accuracy and F1 score, respectively, for 5 rounds of experiments (training and testing), each with a different, random initialization of network weights.</p> <p>For further questions and suggestions, please do not hesitate to contact us.</p> <p> </p> <p><strong>Contact email</strong>: ogomes@gmail.com</p> <p> </p> <p>If you use this dataset in your own work, please cite this DOI: 10.5281/zenodo.5020566</p> <p> </p> <p>Please also cite this paper, which provides additional details about the dataset:</p> <p>Michel Pedro Filippo, Otávio da Fonseca Martins Gomes, Gilson Alexandre Ostwald Pedro da Costa, Guilherme Lucio Abelha Mota. <em>Deep learning semantic segmentation of opaque and non-opaque minerals from epoxy resin in reflected light microscopy images</em>. <strong>Minerals Engineering</strong>, Volume 170, 2021, 107007, https://doi.org/10.1016/j.mineng.2021.107007.</p>
Example image for color based segmentation
<p>With the bright orange pumpkins on a flat background, it is a suitable example for color based segmentation.</p>
Semantic segmentation model of construction waste landfill based on high-resolution satellite images
<p>CWLD_model project shows scripts and instructions on how to use this dataset (<a href="../records/10686118">https://zenodo.org/records/10686118</a>) to train a segmentation model. requirements.txt files provide the libraries you need to run your project. The README.md document details the deployment process and features of each module.</p> <p>You can also visit the GitHub page for scripts and instructions on how to use this dataset for visualizing and plotting basic statistics. The models and the code to execute them are released on <a href="https://github.com/huangleinxidimejd/CWLD_Model">https://github.com/huangleinxidimejd/CWLD_Model</a>.</p> <h2>Training details</h2> <p>The model was trained with two GPUs, an Nvidia GeForce RTX 2080Ti, and the following parameters:</p> <ul> <li>'train_batch_size': 4,</li> <li>'val_batch_size': 4,</li> <li>'train_crop_size': 512,</li> <li>'val_crop_size': 512,</li> <li>'lr': 0.001, # the learning rate used during training. It determines how quickly the model learns from the data</li> <li>'Epoch Times': 200,</li> <li>'gpu': correct,</li> <li>'weight_decay': 5E-4,</li> <li>'Momentum': 0.9,</li> <li>'print_freq': 100,</li> <li>'predict_step': 5,</li> </ul> <h2>usage</h2> <ul> <li>After downloading the dataset from Zenodo, place the train and val files from the Deep Learning Datasets file into the data folder of the CWLD semantic segmentation model.</li> <li>Open: CWLD_ Open the root directory in CWLD_model/dataset/ and start training with the WasteSeg_Train.py file. The modelss module provides five convolutional networks, Improved_DeeplabV3_plus, PSPNet, ResNet, SegNet, and UNet, which can be selected and modified accordingly.</li> <li>The utils package provides a large number of data processing tools to use.</li> <li>The trained model can be predicted from a EvalSeg.py file.</li> </ul>
Raw Metrics and Rankings for "Exploratory Analysis on Pixelwise Image Segmentation Metrics with an Application in Proximal Sensing"
<p>These datasets accompany the article published in <em>Remote Sensing </em>entitled: "Exploratory Analysis on Pixelwise Image Segmentation Metrics with an Application in Proximal Sensing".</p> <p>For each of the three segmentation models presented in the paper (DTSM, SVM and CIVE) two types of datasets are included: </p> <ul> <li><strong>Raw Metrics: </strong>the raw evaluations for each image returned by each of the 12 evaluation metrics. </li> <li><strong>Rankings:</strong> the ranking of each image in the dataset based on its raw evaluation. This dataset has been created by sorting in ascending order the dissimilarity metrics (GCE and HDD) and descending order the similarity metrics (all the other metrics). </li> </ul> <p>The datasets are in Excel (.xlsx) format and can be easily loaded in R and used to reproduce the results presented in the article.</p>
Segmentation of membrane of mouse, sea urchin and human oocytes from transmitted light images
<p>This dataset has been presented in our paper "An interpretable and versatile machine learning approach for oocyte phenotyping", in bioRxiv.</p> <p>It contains images acquired in transmitted light with different settings of mouse and human oocytes and sea urchin eggs, with the corresponding ground-truth of the membrane segmentation. Mouse oocyte images were taken before and during oocyte maturation (meiosis I). Some human oocyte images were taken during oocyte maturation (meiosis I), and some are M-II oocytes just after fertilization. Sea urchin images contains both fertilized and unfertilized eggs.</p> <p> </p>
Segmentation of oocyte zona pellucida in transmitted light images (mouse and human)
<p>This dataset has been presented in our paper "An interpretable and versatile machine learning approach for oocyte phenotyping", in bioRxiv.</p> <p>It contains images acquired in transmitted light with different settings of mouse and human oocytes, with the corresponding ground-truth of the zona pellucida segmentation. Mouse oocyte images were taken before and during oocyte maturation (meiosis I). Some human oocyte images were taken during oocyte maturation (meiosis I), and some are M-II oocytes just after fertilization.</p>
Code and Data from: Segmenting Root Systems in X-Ray Computed Tomography Images Using Level Sets
<p>This record contains code and data for segmentation using a three-dimensional level-set method, written by Amy Tabb in C++. The record also contains two datasets of root systems in media imaged with X-Ray CT, and the results of running the code on those datasets. The code will also perform a pre-processing task in three-dimensional image sets, and a dataset for that purpose is included as well. This work is a companion to the paper : "Segmenting root systems in X-ray computed tomography images using level sets" (WACV 2018) by the authors or this record, and and open-access version of the paper is here -- https://arxiv.org/abs/1809.06398 . The code is also available from Github: https://github.com/amy-tabb/tabb-level-set-segmentation , using a DOI and stable releases https://doi.org/10.5281/zenodo.3344906.</p> <p>Format of the data:</p> <p>Three input datasets are provided; two for the segmentation functionality of the code, and one to test the pre-processing functionality. The two segmentation sets are the same as were used in the paper, and are CassavaDataset, and SoybeanDataset. The pre-processing set is CassavaSlices. The output set for Soybean is SoybeanResultsJul11. The Cassava result set is large, so I broke it into three compressed folders, CassavaResultsJul12_A, _B, _C. _B is the largest, and only contains the results overwritten on the original X-Ray images. Unless your connection to Zenodo is extremely fast, it will be faster to compute the result than to download it.</p> <p> </p> <p> </p><p> </p><p> </p> <p></p> <p></p>
Representative Sample Dataset for Resolution-Agnostic Tissue Segmentation in Whole-Slide Histopathology Images
<p>This is a representative sample from the dataset that was used to develop resolution-agnostic convolutional neural networks for tissue segmentation1 in whole-slide histopathology images.</p> <p>The dataset is composed of two parts: <strong>development set</strong> and <strong>dissimilar set</strong>.</p> <p>Sample images from the development set:</p> <ul> <li>breast_hne_00.tif</li> <li>breast_lymph_node_hne_00.tif</li> <li>tongue_ae1ae3_00.tif</li> <li>tongue_hne_00.tif</li> <li>tongue_ki67_00.tif</li> </ul> <p>Sample images from the dissimilar set:</p> <ul> <li>brain_alcianblue_00.tif</li> <li>cornea_grocott_00.tif</li> <li>kidney_cab_00.tif</li> <li>skin_perls_00.tif</li> <li>uterus_vonkossa_00.tif</li> </ul>
FeM dataset – An iron ore labeled images dataset for segmentation training and testing
<p>This dataset is composed of 81 pairs of correlated images. Each pair contains one image of an iron ore sample acquired through reflected light microscopy (RGB, 24-bit), and the corresponding binary reference image (8-bit), in which the pixels are labeled as belonging to one of two classes: ore (0) or embedding resin (255).</p> <p>The sample came from an itabiritic iron ore concentrate from Quadrilátero Ferrífero (Brazil) mainly composed of hematite and quartz, with little magnetite and goethite. It was classified by size and concentrated with a dense liquid. Then, the fraction -149+105 μm with density greater than 3.2 was cold mounted with epoxy resin and subsequently ground and polished.</p> <p>Correlative microscopy was employed for image acquisition. Thus, 81 fields were imaged on a reflected light microscope with a 10× (NA 0.20) objective lens and on a scanning electron microscope (SEM). In sequence, they were registered, resulting in images of 999×756 pixels with a resolution of 1.05 µm/pixel. Finally, the images from SEM were thresholded to generate the reference images.</p> <p>Further description of this sample and its imaging procedure can be found in the work by Gomes and Paciornik (2012).</p> <p>This dataset was created for developing and testing deep learning models on semantic segmentation tasks. The paper of Filippo et al. (2021) presented a variant of the DeepLabv3+ model that reached mean values of 91.43% and 93.13% for overall accuracy and F1 score, respectively, for 5 rounds of experiments (training and testing), each with a different, random initialization of network weights.</p> <p>For further questions and suggestions, please do not hesitate to contact us.</p> <p> </p> <p><strong>Contact email</strong>: ogomes@gmail.com</p> <p> </p> <p>If you use this dataset in your own work, please cite this DOI: 10.5281/zenodo.5014700</p> <p> </p> <p>Please also cite this paper, which provides additional details about the dataset:</p> <p>Michel Pedro Filippo, Otávio da Fonseca Martins Gomes, Gilson Alexandre Ostwald Pedro da Costa, Guilherme Lucio Abelha Mota. <em>Deep learning semantic segmentation of opaque and non-opaque minerals from epoxy resin in reflected light microscopy images</em>. <strong>Minerals Engineering</strong>, Volume 170, 2021, 107007, https://doi.org/10.1016/j.mineng.2021.107007.</p> <p> </p>
PESO: Prostate Epithelium Segmentation on H&E-stained prostatectomy whole slide images
<p>Large set of whole-slide-images (WSI) of prostatectomy specimens with various grades of prostate cancer (PCa). More information can be found in the corresponding paper: <a href="https://doi.org/10.1038/s41598-018-37257-4">https://doi.org/10.1038/s41598-018-37257-4</a></p> <p>The WSIs in this dataset can be viewed using the open-source software <a href="https://github.com/computationalpathologygroup/ASAP">ASAP</a> or <a href="https://openslide.org/">Open Slide</a>.</p> <p>Due to the large size of the complete dataset, the data has been split up in to multiple archives.</p> <p>The data from the training set:</p> <ul> <li><strong>peso_training_masks.zip: </strong>Training masks (N=62) that have been used to train the main network of our paper. These masks are generated by a trained U-Net on the corresponding IHC slides.</li> <li><strong>peso_training_masks_corrected.zip: </strong>A subset of the color deconvolution masks (N=25) on which manual annotations have been made. Within these regions, stain and other artifacts have been removed.</li> <li><strong>peso_training_colordeconvolution.zip: </strong>Mask files (N=62) containing the P63&CK8/18 channel of the color deconvolution operation. These masks mark all regions that are stained by either P63 or CK8/18 in the IHC version of the slides.</li> <li><strong>peso_training_wsi_{1-6}.zip: </strong>Zip files containing the whole slide images of the training set (N=62). Each archive contains 10 slides, excluding the last which contains 12. These images are exported at a pixel resolution of 0.48mu/pixels. </li> </ul> <p>The data from the test set:</p> <ul> <li><strong>peso_testset_regions.zip: </strong>Collection of annotation XML files with outlines of the test regions. These can be used to view the test regions in more detail using ASAP.</li> <li><strong>peso_testset_png.zip: </strong>Export of the test set regions in PNG format (2500x2500 pixels per region).</li> <li><strong>peso_testset_png_padded.zip: </strong>Export of the test regions in PNG format padded with a 500 pixel wide border (3500x3500 pixels per region). Useful for segmenting pixels at the border of the regions.</li> <li><strong>peso_testset_mapping.csv: </strong>A csv file mapping files from the test set (numbered 1-160) to regions in the xml files. The csv file also contains the label (benign or cancer) for each region.</li> <li><strong>peso_testset_groundtruth_masks.zip: </strong>The ground truth (pixel) masks (N=40) of all regions in the test set. For each pixel in the test set regions, these masks contain the ground truth: 0 for unlabelled, 1 for background and 2 for epithelial tissue.</li> <li><strong>peso_testset_wsi_{1-4}.zip: </strong>Zip files containing the whole slide images of the test set (N=40). Each archive contains 10 slides of the test set. These images are exported at a pixel resolution of 0.48mu/pixels. </li> </ul> <p>This study was financed by a grant from the Dutch Cancer Society (KWF), grant number KUN 2015-7970.</p> <p><strong>If you make use of this dataset please cite both the dataset itself and the corresponding paper: </strong><a href="https://doi.org/10.1038/s41598-018-37257-4">https://doi.org/10.1038/s41598-018-37257-4</a></p> <p><strong>Update July 2021: </strong>We have added the ground truth masks for the test set.</p>
Doodleverse/Segmentation Zoo Res-UNet models for 4-class (water, whitewater, sediment and other) segmentation of Sentinel-2 and Landsat-7/8 1-band NDWI images of coasts.
<p><em><strong>Doodleverse/Segmentation Zoo Res-UNet models for 4-class (water, whitewater, sediment and other) segmentation of Sentinel-2 and Landsat-7/8 1-band NDWI images of coasts.</strong></em></p> <p>These Residual-UNet model data are based on 1-band NDWI images of coasts and associated labels.</p> <p>Models have been created using Segmentation Gym* using the following dataset**: https://doi.org/10.5281/zenodo.7344571</p> <p>Classes: {0=water, 1=whitewater, 2=sediment, 3=other}</p> <p>File descriptions</p> <p>For each model, there are 5 files with the same root name:</p> <p>1. '.json' config file: this is the file that was used by Segmentation Gym* to create the weights file. It contains instructions for how to make the model and the data it used, as well as instructions for how to use the model for prediction. It is a handy wee thing and mastering it means mastering the entire Doodleverse.</p> <p>2. '.h5' weights file: this is the file that was created by the Segmentation Gym* function `train_model.py`. It contains the trained model's parameter weights. It can called by the Segmentation Gym* function `seg_images_in_folder.py`. Models may be ensembled.</p> <p>3. '_modelcard.json' model card file: this is a json file containing fields that collectively describe the model origins, training choices, and dataset that the model is based upon. There is some redundancy between this file and the `config` file (described above) that contains the instructions for the model training and implementation. The model card file is not used by the program but is important metadata so it is important to keep with the other files that collectively make the model and is such is considered part of the model</p> <p>4. '_model_history.npz' model training history file: this numpy archive file contains numpy arrays describing the training and validation losses and metrics. It is created by the Segmentation Gym function `train_model.py`</p> <p>5. '.png' model training loss and mean IoU plot: this png file contains plots of training and validation losses and mean IoU scores during model training. A subset of data inside the .npz file. It is created by the Segmentation Gym function `train_model.py`</p> <p>Additionally, BEST_MODEL.txt contains the name of the model with the best validation loss and mean IoU</p> <p><br> References</p> <p>*Segmentation Gym: Buscombe, D., & Goldstein, E. B. (2022). A reproducible and reusable pipeline for segmentation of geoscientific imagery. Earth and Space Science, 9, e2022EA002332. https://doi.org/10.1029/2022EA002332 See: https://github.com/Doodleverse/segmentation_gym</p> <p>** Buscombe, Daniel. (2022). Images and 4-class labels for semantic segmentation of Sentinel-2 and Landsat RGB, NIR, and SWIR satellite images of coasts (water, whitewater, sediment, other) (v1.0) [Data set]. Zenodo. https://doi.org/10.5281/zenodo.7344571</p> <p> </p>
Doodleverse/Segmentation Zoo Res-UNet models for 4-class (water, whitewater, sediment and other) segmentation of Sentinel-2 and Landsat-7/8 5-band (RGB+NIR+SWIR) images of coasts.
<p>These Residual-UNet model data are based on 5-band RGB+NIR+SWIR (red, green, blue, near-infrared, and short-wave infrared) images of coasts and associated labels.</p> <p> </p> <p>Models have been created using Segmentation Gym* using the following dataset**: <a href="https://doi.org/10.5281/zenodo.7344571">https://doi.org/10.5281/zenodo.7344571 </a></p> <p>Classes: {0=water, 1=whitewater, 2=sediment, 3=other}</p> <p><strong>File descriptions</strong></p> <p>For each model, there are 5 files with the same root name:</p> <p>1. <strong>'.json' </strong>config file: this is the file that was used by Segmentation Gym* to create the weights file. It contains instructions for how to make the model and the data it used, as well as instructions for how to use the model for prediction. It is a handy wee thing and mastering it means mastering the entire Doodleverse.</p> <p>2.<strong> '.h5'</strong> weights file: this is the file that was created by the Segmentation Gym* function `train_model.py`. It contains the trained model's parameter weights. It can called by the Segmentation Gym* function `seg_images_in_folder.py`. Models may be ensembled.</p> <p>3.<strong> '_modelcard.json'</strong> model card file: this is a json file containing fields that collectively describe the model origins, training choices, and dataset that the model is based upon. There is some redundancy between this file and the `config` file (described above) that contains the instructions for the model training and implementation. The model card file is not used by the program but is important metadata so it is important to keep with the other files that collectively make the model and is such is considered part of the model</p> <p>4. <strong> '_model_history.npz'</strong> model training history file: this numpy archive file contains numpy arrays describing the training and validation losses and metrics. It is created by the Segmentation Gym function `train_model.py`</p> <p>5. <strong> '.png'</strong> model training loss and mean IoU plot: this png file contains plots of training and validation losses and mean IoU scores during model training. A subset of data inside the .npz file. It is created by the Segmentation Gym function `train_model.py`</p> <p>Additionally, BEST_MODEL.txt contains the name of the model with the best validation loss and mean IoU</p> <p> </p> <p><strong>References</strong></p> <p>*Segmentation Gym: Buscombe, D., & Goldstein, E. B. (2022). A reproducible and reusable pipeline for segmentation of geoscientific imagery. Earth and Space Science, 9, e2022EA002332. <a href="https://doi.org/10.1029/2022EA002332">https://doi.org/10.1029/2022EA002332</a> See: <a href="https://github.com/Doodleverse/segmentation_gym">https://github.com/Doodleverse/segmentation_gym</a></p> <p>** Buscombe, Daniel. (2022). Images and 4-class labels for semantic segmentation of Sentinel-2 and Landsat RGB, NIR, and SWIR satellite images of coasts (water, whitewater, sediment, other) (v1.0) [Data set]. Zenodo. <a href="https://doi.org/10.5281/zenodo.7344571">https://doi.org/10.5281/zenodo.7344571</a></p>
Images and 4-class labels for semantic segmentation of Sentinel-2 and Landsat RGB, NIR, and SWIR satellite images of coasts (water, whitewater, sediment, other)
<p><em><strong>Images and 4-class labels for semantic segmentation of Sentinel-2 and Landsat RGB, NIR, and SWIR satellite images of coasts (water, whitewater, sediment, other)</strong></em></p> <p><strong>Description</strong></p> <p>579 images and 579 associated labels for semantic segmentation of Sentinel-2 and Landsat RGB satellite images of coasts. The 4 classes are 0=water, 1=whitewater, 2=sediment, 3=other</p> <p>These images and labels have been made using the Doodleverse software package, Doodler*. These images and labels could be used within numerous Machine Learning frameworks for image segmentation, but have specifically been made for use with the Doodleverse software package, Segmentation Gym**.</p> <p>Some (422) of these images and labels were originally included in the Coast Train*** data release, and have been modified from their original by reclassifying from the original classes to the present 4 classes.</p> <p>The label images are a subset of the following data release**** <a href="https://doi.org/10.5281/zenodo.7335647">https://doi.org/10.5281/zenodo.7335647</a></p> <p>Imagery comes from the following 10 sand beach sites:</p> <ol> <li>Duck, NC, Hatteras NC, USA</li> <li>Santa Cruz CA, USA</li> <li>Galveston TX, USA</li> <li>Truc Vert,France</li> <li>Sunset State Beach CA, USA</li> <li>Torrey Pines CA, USA</li> <li>Narrabeen, NSW, Australia</li> <li>Elwha WA, USA</li> <li>Ventura region, CA, USA</li> <li>Klamath region, CA USA</li> </ol> <p>Imagery are a mixture of 10-m Sentinel-2 and 15-m pansharpened Landsat 7, 8, and 9 visible-band imagery of various sizes. Red, Green, Blue, NIR, and SWIR bands only</p> <p><strong>File descriptions</strong></p> <ol> <li>classes.txt, a file containing the class names</li> <li>images.zip, a zipped folder containing the 3-band RGB images of varying sizes and extents</li> <li>nir.zip, a zipped folder containing the corresponding near-infrared (NIR) imagery</li> <li>swir.zip, a zipped folder containing the corresponding shortwave-infrared (SWIR) imagery</li> <li>labels.zip, a zipped folder containing the 1-band label images</li> <li>overlays.zip, a zipped folder containing a semi-transparent overlay of the color-coded label on the image (blue=0=water, red=1=whitewater, yellow=2=sediment, green=3=other)</li> <li>resized_images.zip, RGB images resized to 512x512x3 pixels</li> <li>resized_nir.zip, NIR images resized to 512x512x3 pixels</li> <li>resized_swir.zip, SWIR images resized to 512x512x3 pixels</li> <li>resized_labels.zip, label images resized to 512x512 pixels</li> </ol> <p><strong>References</strong></p> <p>*Doodler: Buscombe, D., Goldstein, E.B., Sherwood, C.R., Bodine, C., Brown, J.A., Favela, J., Fitzpatrick, S., Kranenburg, C.J., Over, J.R., Ritchie, A.C. and Warrick, J.A., 2021. Human‐in‐the‐Loop Segmentation of Earth Surface Imagery. Earth and Space Science, p.e2021EA002085<a href="https://doi.org/10.1029/2021EA002085">https://doi.org/10.1029/2021EA002085</a>. See <a href="https://github.com/Doodleverse/dash_doodler">https://github.com/Doodleverse/dash_doodler.</a></p> <p>**Segmentation Gym: Buscombe, D., & Goldstein, E. B. (2022). A reproducible and reusable pipeline for segmentation of geoscientific imagery. Earth and Space Science, 9, e2022EA002332. <a href="https://doi.org/10.1029/2022EA002332">https://doi.org/10.1029/2022EA002332</a> See: <a href="https://github.com/Doodleverse/segmentation_gym">https://github.com/Doodleverse/segmentation_gym</a></p> <p>***Coast Train data release: Wernette, P.A., Buscombe, D.D., Favela, J., Fitzpatrick, S., and Goldstein E., 2022, Coast Train--Labeled imagery for training and evaluation of data-driven models for image segmentation: U.S. Geological Survey data release, <a href="https://doi.org/10.5066/P91NP87I">https://doi.org/10.5066/P91NP87I</a>. See <a href="https://coasttrain.github.io/CoastTrain/">https://coasttrain.github.io/CoastTrain/ </a>for more information</p> <p>**** Buscombe, Daniel, Goldstein, Evan, Bernier, Julie, Bosse, Stephen, Colacicco, Rosa, Corak, Nick, Fitzpatrick, Sharon, del Jesús González Guillén, Anais, Ku, Venus, Paprocki, Julie, Platt, Lindsay, Steele, Bethel, Wright, Kyle, & Yasin, Brandon. (2022). Images and 4-class labels for semantic segmentation of Sentinel-2 and Landsat RGB satellite images of coasts (water, whitewater, sediment, other) (v1.0) [Data set]. Zenodo. <a href="https://doi.org/10.5281/zenodo.7335647">https://doi.org/10.5281/zenodo.7335647</a></p>
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.