Find research datasets worth reusing

Dataset with segmentations of 117 important anatomical structures in 1228 CT images

Info: This is version 2 of the TotalSegmentator dataset. In 1228 CT images we segmented 117 anatomical structures covering a majority of relevant classes for most use cases. The CT images were randomly sampled from clinical routine, thus representing a real world dataset which generalizes to clinical application. The dataset contains a wide range of different pathologies, scanners, sequences and institutions.Link to a copy of this dataset on Dropbox for much quicker download: <a href="https://www.dropbox.com/scl/fi/oq0fsz8oauory204g8o6f/Totalsegmentator_dataset_v201.zip?rlkey=afnl2ixhqca2ukkf1v9p6jz7p&dl=0">Dropbox Link</a>Overview of differences to v1 of this dataset: <a href="https://github.com/wasserth/TotalSegmentator/blob/master/resources/improvements_in_v2.md">here</a>A small subset of this dataset with only 102 subjects for quick download+exploration can be found here: <a href="https://doi.org/10.5281/zenodo.8367169">here</a>You can find a segmentation model trained on this dataset <a href="https://github.com/wasserth/TotalSegmentator">here</a>. More details about the dataset can be found in the corresponding <a href="https://doi.org/10.1148/ryai.230024">paper</a> (the paper describes v1 of the dataset). Please cite this paper if you use the dataset.This dataset was created by the department of <a href="https://www.unispital-basel.ch/en/radiologie-nuklearmedizin/forschung-radiologie-nuklearmedizin">Research and Analysis at University Hospital Basel</a>.UPDATE: On 2023-10-27 we uploaded version 2.0.1 which fixes broken files.

FIGURE 5 U in Automated segmentation of insect anatomy from micro-CT images using deep learning

FIGURE 5 U-Net implementation. The architecture of the used convolutional neural network (CNN) is an implementation of U-Net. It consists of two parts: two 3×3 convolutions followed by 2×2 max pooling and two 3×3 convolutions followed by 2×2 upconvolutions. Dropout was added to avoid overfitting. As a final step a 1×1 convolution is applied, resulting in an output map with two classes.

FIGURE 6 in Automated segmentation of insect anatomy from micro-CT images using deep learning

FIGURE 6 Network performance evaluation. High true positive rate (TPR) and low false positive rate (FPR) values for training (blue) and testing data (red) indicate the network's high generalizability.

FIGURE 10 in Automated segmentation of insect anatomy from micro-CT images using deep learning

FIGURE 10 Application of pipeline for other insect species. The brain textures of various insect species can be very similar to those of ants, facilitating the prediction by the network even without pretraining on specific insect brain scans. (a) Raw image of wasp head (original 1000 × 1000 px) and (b) its prediction without postprocessing (original 520 × 520 px), indicating satisfactory identification of the borders of the brain area. (c) 2D image of praying mantis head (520 × 520 px) and (d) the prediction of its brain area without postprocessing. Even though the network overpredicts some small pixel islands, it excludes from its prediction areas of the muscles, fibers, and cuticle.

FIGURE 1 in Automated segmentation of insect anatomy from micro-CT images using deep learning

FIGURE 1 Segmentation pipeline overview. (a) Specimens are placed in iodine for staining for 2 weeks and then placed in small vials containing 99% ethanol to prevent them from moving during scanning. (b) The computed tomography (CT) scanner acquires successive X-ray images of the stepwise rotating specimen, and, using a user-defined reference image, automatically reconstructs them to produce orthogonal cross-section stacks that are used for the volume reconstruction of the specimen. (c) Volume rendering for future morphological studies is performed using Amira software. (d) Semiautomated segmentation of the brain volume of each scan (in orange) using the watershed method in Amira. (e) Schematic representation of the U-Net architecture used as the core of the pipeline for the development of a fully automated brain segmentation method. (f) The acquired brain images are used for training after preprocessing augmentation and manual creation of masks. (g) The network's prediction (in yellow) is postprocessed for smoothing out overpredicted areas (in red).

FIGURE 2 in Automated segmentation of insect anatomy from micro-CT images using deep learning

FIGURE 2 Exemplar images of full-body scans from different ant species. Three-dimensional (3D) reconstructed microcomputed tomography (micro-CT) image of (a) Acromyrmex versicolor and (b) Atta texana worker specimens, using volume rendering in Amira. (c) 2D micro-CT full body image of the Atta texana specimen (original 1000 × 1000 px). The brain area is the area with the most uniform pixel density within the whole body in its stained state, which makes it easy to recognize in most high-quality scans.

FIGURE 3 in Automated segmentation of insect anatomy from micro-CT images using deep learning

FIGURE 3 Example of semiautomated brain image segmentation. The brain area (in orange) of an Atta texana ant specimen was segmented using the watershed method in Amira; the 1000 × 1000 × 1000 px 3D image was manually postprocessed by smoothing and cropping oversegmented areas.

FIGURE 9 in Automated segmentation of insect anatomy from micro-CT images using deep learning

FIGURE 9 Prediction of ganglia in the thorax. As the tissue texture in the image is similar to that of the brain, the network accurately predicts other areas of nervous tissue in the organism. The pixel island detection step isolates the brain, but without this step neural tissue can be isolated.

FIGURE 8 3D in Automated segmentation of insect anatomy from micro-CT images using deep learning

FIGURE 8 3D volume of ant brain reconstructed from 2D images (original 520 × 520 px) predicted by the algorithm. 3D reconstructed brain prediction of an Atta texana worker.

FIGURE 7 in Automated segmentation of insect anatomy from micro-CT images using deep learning

FIGURE 7 Pipeline performance demonstrated both for validation (top row) and testing (bottom row) sets. (a, d) Raw images of head of Acromyrmex versicolor and Carebara atoma ant specimens, cropped along the x-y axes. The manually segmented brain areas are indicated in blue. (b, e) Network predictions before postprocessing (in yellow). Areas in yellow dotted circles are pixel islands not connected to the brain area that were overpredicted. (c, f) Predictions after postprocessing (in red). The borders of the predicted areas show good agreement with the manual segmentation in both sets. Note that in overlapping manually and automatically segmented areas in b, c, e, and f, colors appear green or purple.

opencc-by-4.0Nov 2023View details →

CT-LaSTAF continous FLC from AMT29 in 2019

AMT29 / DY110Nov / Dec 2019Atlantic Ocean See also: <a href="https://www.amt-uk.org/">https://www.amt-uk.org/</a>AMT29 in Nov/Dec 2019 was the first field deployment of LabSTAF, consequently not all functions were running as they do in the present instrument, most notably there was no automated acquisition of photosynthetic excitation profiles and associated spectral correction of data.DWM measurements were run automatically, but no PEC correction applied to the data.The instrument was run in continuous FLC mode along the entire transect, the FLC are 'static'.The folder raw_rsf contains the rsf files acquired with timestamp in UTC.Blanks of MQ and 0.2um filtrate were run daily.The folder auxiliary contains the cruise report, the cruisetrack, the troubleshoot log, the blank measurements, and the setting files used. As outlined in the cruise report, many auxiliary parameters (nutrients, particulate absorption, HPLC pigments, 14C-based and O2-based primary production, in line and above water optics) were measured.

opencc-by-4.0Nov 2023View details →

Micro-CT image of cell-populated collagen scaffold in the aqueous environment (contrasted with PTA)

A dataset of the collagen scaffold populated with the 3T3 cells scanned in the aqueous environment using Bruker Skyscan 1276 machine (Bruker, Belgium). Scaffold productionThe processes of obtaining and working with collagen scaffolds were conducted in an isolated environment under sterile conditions. The collagen sponge matrix was manufactured at the Center for Collagen Innovation within the Institute of Regenerative Medicine at Sechenov University and provided to us for experimental purposes. In order to obtain the collagen, the authors utilized animal-derived materials sourced from the tendons of large horned cattle. To do this, the tendons were cleaned of excess tissues, cut into pieces with a thickness of 0.5-1 cm, and sequentially treated for 12 hours in a 0.5 M NaCl solution. Subsequently, the mass was homogenized in a 0.83 M acetic acid solution. The resulting suspension was hydrolyzed with 0.24% pepsin for 2 days, after which 1 M NaOH was added to adjust the pH to 7.5, halting the hydrolysis process. The suspension was precipitated with a 12% NaCl solution, the resulting precipitate was redissolved in 0.02 M acetic acid, and then dialyzed. To obtain collagen porous matrices (sponges), the obtained solution was neutralized using 0.1 M NaOH until a pH of 7-7.5 was reached, and the resulting suspension was lyophilized at -40°C for 2 days.Subsequently, the collagen matrix was cut into cubes with sides measuring 0.5 cm. These cubes were placed in 15 ml test tubes filled with 70% ethyl alcohol for sterilization. The test tubes were then placed on a shaker and left in the refrigerator at +4°C for 24 hours. Afterward, the collagen matrices were removed from the alcohol and rinsed five times with 0.9% NaCl.Following the alcohol rinse to confirm the absence of toxicity, an elution test, adapted following the ISO 10993 protocol, was conducted. To obtain collagen cube extracts, they were incubated in a cell culture medium at a volume of 1 ml per sample for 24 hours at 37°C. The 3T3 cell culture was passaged, with 5000 cells seeded in each well of a 96-well plate. After 24 hours, the cells were treated with extract at a volume of 200 µl per well and left in the incubator at 37°C for 24 hours. The following day, extracts were collected, and AlamarBlue reagent (Invitrogen, Waltham, MA, USA) was added according to the manufacturer's instructions to assess the metabolic activity of the cells. Serial dilutions of sodium dodecyl sulfate (SDS) were used as the positive control. Fluorescence intensity was measured using a Victor Nivo spectrofluorimeter (PerkinElmer, Waltham, Massachusetts, USA) at an excitation wavelength of 530 nm and an emission wavelength of 590 nm.Cell seedingAfter confirming the absence of cytotoxic effects, collagen sponges were seeded with the NIH 3T3 cell line at a density of 50,000 cells per sample (cubes of collagen sponge measuring 0.5 cm per side). Staining techniqueFixed specimens in 10% formalin with PBS were washed after 24 hours with distilled water and after that placed in 3% phosphotungstic acid dissolved in distilled water for 24 hours and kept on the rotary shaker at room temperature. After staining, samples were washed and stored in distilled water at 5 °C. Image acquisition and reconstructionA plastic tube filled with distilled water containing the contrasted sample was placed on the sample holder in a SkyScan 1276 micro-CT (Bruker, Kontich, Belgium) and were scanned at 3 μm voxel resolution with 70 kV voltage and 200 uA source power and an aluminum filter with 1 mm of thickness. The rotation was set to 360° around the vertical axis of the sample, with two middle frames for each 0.2° angle step.After scanning, the data were reconstructed using Bruker's NRecon software. During reconstruction, the ring artifact reduction value was set to 20% and the beam hardening correction value to 30%. After that, samples were exported as a series of 16-bit TIFF images which could be opened in the specialized software.

opencc-by-4.0Nov 2023View details →

The supplemental data for the paper: "Methodology of generation of CFD meshes and 4D shape reconstruction of coronary arteries from patient-specific dynamic CT"

The supplemental data for the paper: "Methodology of generation of CFD meshes and  4D shape reconstruction of coronary arteries from patient-specific dynamic CT"A video file (minimum play resolution is HD to see the mesh) showing the movement of the LCA throughout the heart cycle and .STL files for 10--100% (increment of 10\%) of the heart cycle phase.

Figure 14 in Mastiglanis durantoni from French Guyana, a second species in the genus (Siluriformes: Heptapteridae), with a CT scan survey of phylogenetically-relevant characters

Figure 14. – Schematic illustrations of the occipital process and part of posterior cranial fontanel in species of Mastiglanis. A: Mastiglanis durantoni, MZUSP 118119, paratype; B-G: M. asopos, B: MZUSP 93307 (Rio Negro basin), C: MZUSP 81411 (Rio Negro basin), D: MZUSP 86958 (Rio Preto da Eva basin), E-G: MZUSP 97150 (Rio Xingu basin). Scale bars = 1 mm.

Figure 13 in Mastiglanis durantoni from French Guyana, a second species in the genus (Siluriformes: Heptapteridae), with a CT scan survey of phylogenetically-relevant characters

Figure 13. – CT scan image of posterior portion of skull and anterior part of vertebral column of Mastiglanis durantoni, MZUSP 118118, paratype, ventral view. Anterior to top. Abbreviations: bo, basioccipital; ex, exoccipital; fr, frontal; pa, parasphenoid; par, parapophysis; plr, pleural rib; pro, prootic; pt, pterotic; ptr, pterosphenoid; pts, posttemporo-supracleithrum; sph, sphenotic; tp4a, anterior ramus of transverse process of vertebra 4; tp4p, posterior ramus of transverse process of vertebra 4; tp5, transverse process of vertebra 5; tr, tripus; trs, transscapular process; vc5-7, vertebral centra 5 to 7.

Figure 12 in Mastiglanis durantoni from French Guyana, a second species in the genus (Siluriformes: Heptapteridae), with a CT scan survey of phylogenetically-relevant characters

Figure 12. – CT scan image of suspensorium of Mastiglanis durantoni, MZUSP 118118, paratype, lateral view. Anterior to left. Abbreviations: ent, entopterygoid; hy, hyomandibula; io, interopercle; mt, metapterygoid; op, opercle; po, preopercle; qu, quadrate; sbpo, subpreopercle; spo, suprapreopercle.

Figure 9 in Mastiglanis durantoni from French Guyana, a second species in the genus (Siluriformes: Heptapteridae), with a CT scan survey of phylogenetically-relevant characters

Figure 9. – Live specimen of Mastiglanis durantoni n. sp, lateral view, French Guiana (photo by P.Y. Le Bail). Specimen not preserved.

Figure 6 in Mastiglanis durantoni from French Guyana, a second species in the genus (Siluriformes: Heptapteridae), with a CT scan survey of phylogenetically-relevant characters

Figure 6. – CT scan image of premaxilla of Mastiglanis durantoni, MZUSP 118118, paratype. A: Dorsal view; B: Ventral view. Anterior to top.

Figure 7 in Mastiglanis durantoni from French Guyana, a second species in the genus (Siluriformes: Heptapteridae), with a CT scan survey of phylogenetically-relevant characters

Figure 7. – CT scan image of lower jaw of Mastiglanis durantoni, MZUSP 118118, paratype, lateral view, left side, anterior to left. Abbreviations: aa, anguloarticular; den, dentary; lsp, opening for latero-sensory pore; tr, trabeculae.

Figure 5 in Mastiglanis durantoni from French Guyana, a second species in the genus (Siluriformes: Heptapteridae), with a CT scan survey of phylogenetically-relevant characters

Figure 5. – CT scan image of anterior portion of neurocranium of Mastiglanis durantoni, MZUSP 118118, paratype, dorsal view. Abbreviations: fr, frontal; le, lateral ethmoid; me, mesethmoid.