Find research datasets worth reusing

opencc-by-4.0Oct 2022View details →

Datasets of "Influence of contrast and texture based image modifications on the performance and attention shift of U-Net models for brain tissue segmentation" Part 9 of 14

This dataset is part of the work <a href="https://www.frontiersin.org/articles/10.3389/fnimg.2022.1012639/full">https://www.frontiersin.org/articles/10.3389/fnimg.2022.1012639/full</a>. This is the ninth part of 14 parts of the full dataset (9/14). It contains 3 sets of simulated T1 weighted brain volumes in 3 simulated scanning sequences of spin-echo. The parameters of simulated scanning sequences are respectively repetition time (TR) = 600ms, 700ms,  800ms, and echo time (TE) = 15ms. Under each simulated scanning sequence, there are 500 brain volumes. The segmentation labels for each tissue are contained in the first part which you may find at <a href="https://zenodo.org/record/7294916">https://zenodo.org/record/7294916</a> The simulation process of this dataset involves two processes. The first is to simulate one brain under different simulated scanning sequences. For this, we use BrainWeb <a href="https://brainweb.bic.mni.mcgill.ca/cgi/bw/submit_request">https://brainweb.bic.mni.mcgill.ca/cgi/bw/submit_request</a>. In the custom setting, we use spin-echo and apply image artifact the same as the default setting of this page. The second process is to transform each simulated brain from BrainWeb to different anatomical shapes. We use Human Connectome Project (HCP) 1200 subject data <a href="https://www.humanconnectome.org/study/hcp-young-adult">https://www.humanconnectome.org/study/hcp-young-adult</a> and randomly select 500 brains as anatomical references. Other details of this dataset can be found at <a href="https://www.frontiersin.org/articles/10.3389/fnimg.2022.1012639/full">https://www.frontiersin.org/articles/10.3389/fnimg.2022.1012639/full</a> where the details of the data construction are discussed. All parts of the whole dataset can be found at: Part 1: <a href="https://zenodo.org/record/7294916">https://zenodo.org/record/7294916</a> Part 2: <a href="https://zenodo.org/record/7389550">https://zenodo.org/record/7389550</a> Part 3: <a href="https://zenodo.org/record/7390382">https://zenodo.org/record/7390382</a> Part 4: <a href="https://zenodo.org/record/7390741">https://zenodo.org/record/7390741</a> Part 5: <a href="https://zenodo.org/record/7391205">https://zenodo.org/record/7391205</a> Part 6: <a href="https://zenodo.org/record/7393060">https://zenodo.org/record/7393060</a> Part 7: <a href="https://zenodo.org/record/7393174">https://zenodo.org/record/7393174</a> Part 8: <a href="https://zenodo.org/record/7393347">https://zenodo.org/record/7393347</a> Part 9: <a href="https://zenodo.org/record/7394250">https://zenodo.org/record/7394250</a> Part 10: <a href="https://zenodo.org/record/7394667">https://zenodo.org/record/7394667</a> Part 11: <a href="https://zenodo.org/record/7394939">https://zenodo.org/record/7394939</a> Part 12: <a href="https://zenodo.org/record/7395031">https://zenodo.org/record/7395031</a> Part 13: <a href="https://zenodo.org/record/7395620">https://zenodo.org/record/7395620</a> Part 14: <a href="https://zenodo.org/record/7395622">https://zenodo.org/record/7395622</a>

opencc-by-4.0Dec 2022View details →

Retinal Fundus Multi-Disease Image Dataset (RFMiD) 2.0

Retinal Fundus Multi-disease Image Dataset (RFMiD 2.0) is an auxiliary dataset to our previously published dataset. RFMiD 2.0 is a more challenging dataset to research society to develop the computer-based disease diagnosis system. Diabetic Retinopathy, cataracts, and refractive error in the eye are leading diseases that may lead to permanent vision loss more frequently. Therefore, developing an AI-based model to classify these diseases is useful for ophthalmologists. This dataset consists of 860 images of frequently and rarely observed 51 diseases. However, some images are labeled with multiple diseases. This dataset is useful for the research and development of AI-based medical healthcare systems in ophthalmology.

opencc-by-4.0Jan 2023View details →

StreetScouting dataset: A Street-Level Image dataset for finetuning and applying custom object detectors for urban feature selection

The dataset consists of two .zip files. The first .zip file named "annotated dataset" contains a folder named “annotated dataset" with annotated street images. It consists of the folder “images” that has 763 image files. The image format is PNG. 432 images have dimensions of 1080 x 2160 and 331 have dimensions of 866 x 2400. The filenames are random uuids. The “annotated dataset” folder also contains the annotations in the file “coco_annotations.json”. Annotations are provided in COCO format. Table 1 shows the total number of annotated objects per class. <table align="center" summary="Total number of annotated objects per class"> <caption>Table 1. Total number of annotated objects per class</caption> <thead> <tr> <th scope="col">Class</th> <th scope="col">Annotated Objects</th> </tr> </thead> <tbody> <tr> <td>Tree</td> <td>1922</td> </tr> <tr> <td>Waste Bin</td> <td>223</td> </tr> <tr> <td>Recycling Bin</td> <td>181</td> </tr> <tr> <td>Lighting Pole</td> <td>716</td> </tr> <tr> <td>Shop Storefront</td> <td>628</td> </tr> </tbody> </table> The second .zip file is named "routes" and contains a folder named “routes” with consecutive frames of four different driving routes in the city of Thessaloniki and their corresponding GPS signal. So the folder “routes” contains 4 folders in the following format “VID_<YYYYMMDD>_<HHmmSS>” where Y denotes digits for year, M denotes digits for month, D denotes digits for day, H denotes digits for hour, m denotes digits for minutes and S denotes digits for seconds. Not the filename represents the start of the collection sequence. All street data was collected in 2022. Each route folder has the “images” folder which contains the consecutive street image data. Image data in this folder is in JPEG format. Each filename in ‘images’ has the frame_<id>.jpg format where id denotes the order of the frame. Table 2 shows more details regarding the number of frames and frame dimension of the driving routes. <table align="center" summary="Total number of annotated objects per class"> <caption>Table 2. Total number frames and frame dimensions for each of the routes</caption> <thead> <tr> <th scope="col">Route Name</th> <th scope="col">Frames Number</th> <th scope="col">Frame Dimension</th> <th scope="col">Route duration</th> </tr> </thead> <tbody> <tr> <td>VID_20220617_111456</td> <td>41,650</td> <td>1080 x 2160</td> <td>1h, 9m, 26s</td> </tr> <tr> <td>VID_20220210_112926</td> <td>23.035</td> <td>866 x 2400</td> <td>38m, 26s</td> </tr> <tr> <td>VID_20220209_114831</td> <td>18.000</td> <td>1080 x 2160</td> <td>30m, 3s</td> </tr> <tr> <td>VID_20220209_123323</td> <td>18.273</td> <td>1080 x 2160</td> <td>30m, 30s</td> </tr> </tbody> </table> Each route folder contains a “gps.json” file which contains latitude and longitude information for each frame. This file is essentially a JSON list of objects that each object contains the “frame_name” attribute and the corresponding “coordinates” object which contains the “latitude” and "longitude" attributes.

Imaging dataset 01 for mtFociCounter

This is the first dataset associated with the updated manuscript on mtFociCounter. Images are Spinning Disk Confocal Images of unsorted 3t3 NIH mouse fibroblasts stably expressing mitochondrially targeted dsRed, immunofluorescence against dsDNA (mitochondrial nucleoids) and AlexaFluor 647, and stained nuclei with Hoechst. It contains all data necessary to reproduce the analysis of 3t3 WT sampling (sample-to-sample comparison), as described in the manuscript. "samples" from the same day are 13mm coverslips processed in parallel (seeding of cells, fixation, immunofluorescence) and mounted on the same glass slide (#1 on left, #2 centre, #3 right). The dataset contains the 3t3 WT data from the following acquisition dates: 20220603_sample2 20220603_sample3 20220706_sample2 20220706_sample3 20220708_sample1 20220708_sample2 20220708_sample3 20221022_sample2 20221022_sample3 For details of the sampling procedure, please refer to the accompanying manuscript, which will soon be made available on BioRxiv.

Imaging dataset 03 for mtFociCounter

This is the third dataset associated with the updated manuscript on mtFociCounter. Images are Spinning Disk Confocal Images of unsorted 3t3 NIH mouse fibroblasts stably expressing mitochondrially targeted dsRed, immunofluorescence against dsDNA (mitochondrial nucleoids) and AlexaFluor 647, and nuclei stained with Hoechst. Together with dataset_01, it contains all data necessary to reproduce the analysis of 3t3 WT cells, as described in the manuscript. "samples" from the same day are 13mm coverslips processed in parallel (seeding of cells, fixation, immunofluorescence and imaging) and mounted on the same glass slide (#1 on left, #2 centre, #3 right), whereas different dates can be considered biological replicates, of which there are 19 in total (with dataset 01). Please beware, there is some redundancy with dataset_02, and we recommend combining dataset_01 with dataset_03 for further analysis. This dataset_03 contains the 3t3 WT data from the following acquisition dates: 20221025_sample1 20221026_sample2 20221029_sample1 20221029_sample2 20221110_sample2 20221210_sample1 20221211_sample3 20221212_sample3 20221217_sample2 20221220_sample2 20221221_sample3 For further details on the experimental procedure, please refer to the accompanying manuscript, which will soon be made available on BioRxiv.

Imaging dataset 02 for mtFociCounter

This is the second dataset associated with the updated manuscript on mtFociCounter. Images are Spinning Disk Confocal Images of unsorted 3t3 NIH mouse fibroblasts stably expressing mitochondrially targeted dsRed, immunofluorescence against dsDNA (mitochondrial nucleoids) and AlexaFluor 647, and stained nuclei with Hoechst. It contains all data necessary to reproduce the testing of mtFociCounter (no primary control; rho0 control; manual cell segmentation), as described in the manuscript. Samples were processed and imaged in parallel, on the same date, in biological replicates (different date). The dataset contains the 3t3 WT data from the following acquisition dates: no primary antibody (no1) vs. with primary antibody against dsDNA (nucleoids): 20221022_no1 vs. wt 20221025_no1 vs. wt 20221029_no1 vs. wt for 3t3 cells depleted of mtDNA (rho0) vs. normal 3t3 cells with mtDNA (wt): 20221223: sample_1 & sample_2 rho0 vs. sample_3 wt for 3t3 WT cells, same raw-data, manually segmented single cells on two independent days: 20221022 -> segmentation on 20221102 vs. segmentation on 20221103 20221029 -> segmentation on 20221102 vs. segmentation on 20221104 For further details of the exact experimental procedure, please refer to the accompanying manuscript, which will soon be made available on BioRxiv.

Dataset of construction and demolition waste images: aerated autoclaved concrete (AAC), asphalt, ceramics, and concrete

Image subsets: the dataset of images (RGB) of CDW materials (aerated autoclaved concrete (AAC), asphalt, ceramics, and concrete) cropped to 200x200 px. The images are annotated and split into testing and training datasets for the purposes of machine-learning models' training. Whole CDW fragments: images used for validation of algorithms - whole fragments placed on contrast background.

opencc-by-4.0Mar 2023View details →

Image dataset for training of an insect classification model

This version is deprecated! Please use the updated <a href="https://doi.org/10.5281/zenodo.8325383">Insect Detect - insect classification dataset v2</a> with more images and classes. This dataset contains images of various insects and some other arthropods, sitting on or flying above an artificial flower platform. All images were automatically recorded with the <a href="https://maxsitt.github.io/insect-detect-docs/">Insect Detect DIY camera trap</a>, a hardware combination of the Luxonis OAK-1, Raspberry Pi Zero 2 W and PiJuice Zero pHAT for automated insect monitoring (<a href="https://doi.org/10.1101/2023.12.05.570242">bioRxiv preprint</a>).This classification dataset contains the cropped bounding boxes, exported from the <a href="https://universe.roboflow.com/maximilian-sittinger/insect_detect_detection/dataset/6">Insect_Detect_detection</a> dataset together with 290 new images of Episyrphus balteatus.<h2>Classes</h2>The following classes were annotated in this dataset:<ul><li>wasp (mostly Vespula sp.)</li><li>hbee (Apis mellifera)</li><li>fly (mostly Brachycera)</li><li>hovfly (various Syrphidae, e.g. Eupeodes corollae, Scaeva pyrastri)</li><li>episyr_balt (Episyrphus balteatus)</li><li>other (all Arthropods with insufficient occurences, e.g. various Hymenoptera, true bugs, beetles)</li><li>shadow (shadows of the recorded insects)</li></ul>View the <a href="https://universe.roboflow.com/maximilian-sittinger/insect_detect_classification/health">Health Check</a> for more info on class balance.<h2>Deployment</h2>You can use this dataset as starting point to train your own insect classification models. Check the <a href="https://maxsitt.github.io/insect-detect-docs/modeltraining/train_classification/">model training instructions</a> for more information.To deploy the image classification model (ONNX format) on your PC for fast CPU inference, follow the provided <a href="https://maxsitt.github.io/insect-detect-docs/deployment/classification/">Step by Step instructions</a>. Open source Python scripts to deploy the trained model can be found at the <a href="https://github.com/maxsitt/insect-detect-ml">insect-detect-ml GitHub repo</a>.

opencc-by-4.0Mar 2023View details →

Image dataset for training of an insect detection model for the Insect Detect DIY camera trap

This dataset contains images of an artifical flower platform with different insects sitting on it or flying above it. All images were automatically recorded with the <a href="https://maxsitt.github.io/insect-detect-docs/">Insect Detect DIY camera trap</a>, a hardware combination of the Luxonis OAK-1, Raspberry Pi Zero 2 W and PiJuice Zero pHAT for automated insect monitoring (<a href="https://doi.org/10.1101/2023.12.05.570242">bioRxiv preprint</a>).<h2>Classes</h2>The following object classes were annotated in this dataset:<ul><li>wasp (mostly Vespula sp.)</li><li>hbee (Apis mellifera)</li><li>fly (mostly Brachycera)</li><li>hovfly (various Syrphidae, e.g. Episyrphus balteatus)</li><li>other (all Arthropods with insufficient occurences, e.g. various Hymenoptera, true bugs, beetles)</li><li>shadow (shadows of the recorded insects)</li></ul>View the <a href="https://universe.roboflow.com/maximilian-sittinger/insect_detect_detection/health">Health Check</a> for more info on class balance.<h2>Versions</h2><ul><li><a href="https://universe.roboflow.com/maximilian-sittinger/insect_detect_detection/dataset/4">v4 insect_detect_416_1class</a><ul><li>squashed to square (aspect ratio 1:1)</li><li>downscaled to 416x416 pixel</li><li>all classes merged into one class ("insect")</li></ul></li><li><a href="https://universe.roboflow.com/maximilian-sittinger/insect_detect_detection/dataset/5">v5 insect_detect_raw_4K</a><ul><li>original images in 4K resolution (3840x2160 pixel)</li></ul></li><li><a href="https://universe.roboflow.com/maximilian-sittinger/insect_detect_detection/dataset/7">v7 insect_detect_320_1class</a><ul><li>squashed to square (aspect ratio 1:1)</li><li>downscaled to 320x320 pixel</li><li>all classes merged into one class ("insect")</li></ul></li></ul><h2>Deployment</h2>You can use this dataset as starting point to train your own insect detection models. Check the <a href="https://maxsitt.github.io/insect-detect-docs/modeltraining/train_detection/">model training instructions</a> for more information.Open source Python scripts to deploy the trained models can be found at the <a href="https://github.com/maxsitt/insect-detect">insect-detect GitHub repo</a>.

opencc-by-4.0Mar 2023View details →

Dataset of phantom images captured using a multi-flash and a cross polarised probe.

Data captured using two micro camera based imaging probes. One uses a multi-flash technique to eliminate specular reflections. The other uses a cross polarisation technique to overcome specular reflections. Images of embedded features in a multi-layer tissue mimicking phantom are captured using different illumination wavelengths at different depths.

Multi-topography dataset for wind turbine detection from remote sensing image

The land remote sensing wind turbine dataset has 1270 remote sensing images and contains 4459 individual wind turbines.  The images are taken over a large time span and contain remote sensing images of the same wind farm at different times. The dataset has both YOLO and VOC tagging formats. The dataset can be divided into five categories based on the land background: sandy land, forest land, grassland, snow land, and wasteland. Rich land background can improve the robustness of the model, and different marker formats and a large number of wind turbine individuals can meet the object detection of different models. Affected by the size of different power wind turbines, the multi-angle imaging characteristics of remote sensing satellites, the different solar radiation angles in different seasons and vegetation shading, wind turbines show large differences in the images. The image features of the wind turbine shadow are more obvious than those of the wind turbine body, so in order for the detection model to better identify the wind turbine, we label the wind turbine body and the wind turbine shadow as a whole when using the labelImg tool for labeling the wind turbine target.

Training dataset for spectral compressive imaging in DL4sSR

Training dataset should be placed as \DL4sSR\3SCI\26train_256_enhanced.h5 Benchmark: <a href="https://github.com/JiangHe96/DL4sSR">https://github.com/JiangHe96/DL4sSR</a> Reference: J. He, Q. Yuan, J. Li, Y. Xiao, D. Liu, H. Shen, and L. Zhang, "Spectral super-resolution meets deep learning: achievements and challenges," Information Fusion, 2023.

Low-dose Computed Tomography Perceptual Image Quality Assessment Grand Challenge Dataset (MICCAI 2023)

Image quality assessment (IQA) is extremely important in computed tomography (CT) imaging, since it facilitates the optimization of radiation dose and the development of novel algorithms in medical imaging, such as restoration. In addition, since an excessive dose of radiation can cause harmful effects in patients, generating high- quality images from low-dose images is a popular topic in the medical domain. However, even though peak signal-to-noise ratio (PSNR) and structural similarity index measure (SSIM) are the most widely used evaluation metrics for these algorithms, their correlation with radiologists’ opinion of the image quality has been proven to be insufficient in previous studies, since they calculate the image score based on numeric pixel values (1-3). In addition, the need for pristine reference images to calculate these metrics makes them ineffective in real clinical environments, considering that pristine, high-quality images are often impossible to obtain due to the risk posed to patients as a result of radiation dosage. To overcome these limitations, several studies have aimed to develop a no-reference novel image quality metric that correlates well with radiologists’ opinion on image quality without any reference images (2, 4, 5). Nevertheless, due to the lack of open-source datasets specifically for CT IQA, experiments have been conducted with datasets that differ from each other, rendering their results incomparable and introducing difficulties in determining a standard image quality metric for CT imaging. Besides, unlike real low-dose CT images with quality degradation due to various combinations of artifacts, most studies are conducted with only one type of artifact (e.g., low-dose noise (6-11), view aliasing (12), metal artifacts (13), scattering (14-16), motion artifacts (17-22), etc.). Therefore, this challenge aims to 1) evaluate various NR-IQA models on CT images containing complex noise/artifacts, 2) to compare their correlations with scores produced by radiologists, and 3) to grant insights into the determination of the best-performing metric of CT imaging in terms of correlating with the perception of radiologists’. Furthermore, considering that low-dose CT images are achieved by reducing the number of projections per rotation and by reducing the X-ray current, the combination of two major artifacts, namely the sparse view streak and noise generated by these methods, is dealt with in this challenge so that the best-performing IQA model applicable in real clinical environments can be verified.   Funding Declaration: This research was partly supported by Institute of Information & communications Technology Planning & Evaluation (IITP) grant funded by the Korea government(MSIT) (No.RS-2022-00155966, Artificial Intelligence Convergence Innovation Human Resources Development (Ewha Womans University)), and by the National Research Foundation of Korea (NRF-2022R1A2C1092072), and by the Korea Medical Device Development Fund grant funded by the Korea government (the Ministry of Science and ICT, the Ministry of Trade, Industry and Energy, the Ministry of Health & Welfare, the Ministry of Food and Drug Safety) (Project Number: 1711174276, RS-2020-KD000016).   References: <ol> <li>Lee W, Cho E, Kim W, Choi J-H. Performance evaluation of image quality metrics for perceptual assessment of low-dose computed tomography images. Medical Imaging 2022: Image Perception, Observer Performance, and Technology Assessment: SPIE, 2022.</li> <li>Lee W, Cho E, Kim W, Choi H, Beck KS, Yoon HJ, Baek J, Choi J-H. No-reference perceptual CT image quality assessment based on a self-supervised learning framework. Machine Learning: Science and Technology 2022.</li> <li>Choi D, Kim W, Lee J, Han M, Baek J, Choi J-H. Integration of 2D iteration and a 3D CNN-based model for multi-type artifact suppression in C-arm cone-beam CT. Machine Vision and Applications 2021;32(116):1-14.</li> <li>Pal D, Patel B, Wang A. SSIQA: Multi-task learning for non-reference CT image quality assessment with self-supervised noise level prediction. 2021 IEEE 18th International Symposium on Biomedical Imaging (ISBI): IEEE, 2021; p. 1962-1965.</li> <li>Mittal A, Moorthy AK, Bovik AC. No-reference image quality assessment in the spatial domain. IEEE Trans Image Process 2012;21(12):4695-4708. doi: 10.1109/TIP.2012.2214050</li> <li>Lee J-YK, Wonjin; Lee, Yebin; Lee, Ji-Yeon; Ko, Eunji; Choi, Jang-Hwan. Unsupervised Domain Adaptation for Low-dose Computed Tomography Denoising. IEEE Access 2022.</li> <li>Jeon S-Y, Kim W, Choi J-H. MM-Net: Multi-frame and Multi-mask-based Unsupervised Deep Denoising for Low-dose Computed Tomography. IEEE Transactions on Radiation and Plasma Medical Sciences 2022.</li> <li>Kim W, Lee J, Kang M, Kim JS, Choi J-H. Wavelet subband-specific learning for low-dose computed tomography denoising. PloS one 2022;17(9):e0274308.</li> <li>Han M, Shim H, Baek J. Low-dose CT denoising via convolutional neural network with an observer loss function. Med Phys 2021;48(10):5727-5742. doi: 10.1002/mp.15161</li> <li>Kim B, Shim H, Baek J. Weakly-supervised progressive denoising with unpaired CT images. Med Image Anal 2021;71:102065. doi: 10.1016/j.media.2021.102065</li> <li>Wagner F, Thies M, Gu M, Huang Y, Pechmann S, Patwari M, Ploner S, Aust O, Uderhardt S, Schett G, Christiansen S, Maier A. Ultralow-parameter denoising: Trainable bilateral filter layers in computed tomography. Med Phys 2022;49(8):5107-5120. doi: 10.1002/mp.15718</li> <li>Kim B, Shim H, Baek J. A streak artifact reduction algorithm in sparse-view CT using a self-supervised neural representation. Med Phys 2022. doi: 10.1002/mp.15885</li> <li>Kim S, Ahn J, Kim B, Kim C, Baek J. Convolutional neural network-based metal and streak artifacts reduction in dental CT images with sparse-view sampling scheme. Med Phys 2022;49(9):6253-6277. doi: 10.1002/mp.15884</li> <li>Bier B, Berger M, Maier A, Kachelrieß M, Ritschl L, Müller K, Choi JH, Fahrig R. Scatter correction using a primary modulator on a clinical angiography Carm CT system. Med Phys 2017;44(9):e125-e137.</li> <li>Maul N, Roser P, Birkhold A, Kowarschik M, Zhong X, Strobel N, Maier A. Learning-based occupational x-ray scatter estimation. Phys Med Biol 2022;67(7). doi: 10.1088/1361-6560/ac58dc</li> <li>Roser P, Birkhold A, Preuhs A, Syben C, Felsner L, Hoppe E, Strobel N, Kowarschik M, Fahrig R, Maier A. X-Ray Scatter Estimation Using Deep Splines. IEEE Trans Med Imaging 2021;40(9):2272-2283. doi: 10.1109/TMI.2021.3074712</li> <li>Maier J, Nitschke M, Choi JH, Gold G, Fahrig R, Eskofier BM, Maier A. Rigid and Non-Rigid Motion Compensation in Weight-Bearing CBCT of the Knee Using Simulated Inertial Measurements. IEEE Trans Biomed Eng 2022;69(5):1608-1619. doi: 10.1109/TBME.2021.3123673</li> <li>Choi JH, Maier A, Keil A, Pal S, McWalter EJ, Beaupré GS, Gold GE, Fahrig R. Fiducial markerbased correction for involuntary motion in weightbearing Carm CT scanning of knees. II. Experiment. Med Phys 2014;41(6Part1):061902.</li> <li>Choi JH, Fahrig R, Keil A, Besier TF, Pal S, McWalter EJ, Beaupré GS, Maier A. Fiducial markerbased correction for involuntary motion in weightbearing Carm CT scanning of knees. Part I. Numerical modelbased optimization. Med Phys 2013;40(9):091905.</li> <li>Berger M, Muller K, Aichert A, Unberath M, Thies J, Choi JH, Fahrig R, Maier A. Marker-free motion correction in weight-bearing cone-beam CT of the knee joint. Med Phys 2016;43(3):1235-1248. doi: 10.1118/1.4941012</li> <li>Ko Y, Moon S, Baek J, Shim H. Rigid and non-rigid motion artifact reduction in X-ray CT using attention module. Med Image Anal 2021;67:101883. doi: 10.1016/j.media.2020.101883</li> <li>Preuhs A, Manhart M, Roser P, Hoppe E, Huang Y, Psychogios M, Kowarschik M, Maier A. Appearance Learning for Image-Based Motion Estimation in Tomography. IEEE Trans Med Imaging 2020;39(11):3667-3678. doi: 10.1109/TMI.2020.3002695</li> </ol>

Reconstructing the Image Scanning Microscopy Dataset: an Inverse Problem

It contains the simulation scripts and the data acquired and analyzed to produce the manuscript "Reconstructing the Image Scanning Microscopy Dataset: an Inverse Problem" (DOI: <a href="https://doi.org/10.1088/1361-6420/accdc5"> 10.1088/1361-6420/accdc5</a>).

openother-openApr 2023View details →

Dataset: Imaging the columnar functional organization of human area MT+ to axis-of-motion stimuli using VASO at 7 Tesla

First release upon acceptance of the paper.

Co-manipulation of soft-materials estimating deformation from depth images [Dataset]

The dataset and training results used in the paper "Co-manipulation of soft-materials estimating deformation from depth images" G. Nicola, E. Villagrossi, N.Pedrocchi submitted at Robotics and Computer Integrated Manufacturing.

opencc-by-4.0May 2023View details →

Line Segment in Document Images Datasets

Those datasets are related to the accepted article to ICDAR 2023: "Linear Object Detection in Document Images using Multiple Object Tracking" by Bernet et al. The official github repository is : https://github.com/EPITAResearchLab/bernet.23.icdar .png ground truth are labelled images where white pixels correspond to the background. .csv ground truth are using the following format: ``` x,y,w,h       # First line is the size of the interest area in the image x1,y1,x2,y2   # Following lines are the coordinates of the extremities of the line ... xn,yn,xn+1,yn+1 ```

opencc-by-4.0May 2023View details →

A large expert-curated cryo-EM image dataset for machine learning protein particle picking

Cryo-electron microscopy (cryo-EM) is a powerful technique for determining the structures of biological macromolecular complexes. Picking single-protein particles from cryo-EM micrographs is a crucial step in reconstructing protein structures. However, the widely used template-based particle picking process is labor-intensive and time-consuming. Though machine learning and artificial intelligence (AI) based particle picking can potentially automate the process, its development is hindered by lack of large, high-quality labelled training data. To address this bottleneck, we present CryoPPP, a large, diverse, expert-curated cryo-EM image dataset for protein particle picking and analysis. It consists of labelled cryo-EM micrographs (images) of 34 representative protein datasets selected from the Electron Microscopy Public Image Archive (EMPIAR). The dataset is 2.6 terabytes and includes 9,893 high-resolution micrographs with labelled protein particle coordinates. The labelling process was rigorously validated through 2D particle class validation and 3D density map validation with the gold standard. The dataset is expected to greatly facilitate the development of both AI and classical methods for automated cryo-EM protein particle picking.

opencc-by-4.0May 2023View details →

image classification dataset on carbon fiber reinforcement quality control

Image classification dataset on carbon fiber quality control. To represent a practical quality control problem, a dataset was generated using carbon plain weave with a grammage of 200g/m². Pieces of the weave, measuring 300x300 mm², were cut using a CNC cutter table. Two such pieces were stacked, with a binder applied between them for shape stability after the forming process. The formed stacks were then scanned using a high-resolution camera mounted on a robotic arm, resulting in 500 images of the textiles' surfaces in three-dimensional shape. The images were then cropped to 341x384 pixels patches and transformed to grayscale. Each patch was classified into one of three classes: normal textile, gap, or fold. Images that were blurred, out of focus, or had bad contrast were sorted out. The dataset has not yet been released, and will be available upon the acceptance of the document.