Data for: A hybrid deep learning model enables automated multi-organ segmentation of whole-body zebrafish micro-computed tomography

Description

This dataset accompanies the manuscript "A hybrid deep learning model enables automated multi-organ segmentation of whole-body zebrafish micro-computed tomography" (Reza et al., submitted to Scientific Reports). It contains a whole-body micro-computed tomography (micro-CT) volume of one zebrafish (Danio rerio), with expert manual annotations of four anatomical structures: the brain, heart, left eye and right eye. The specimen was raised in the Penn State Zebrafish Functional Genomics Core, euthanized in MS-222 (Tricaine-S), fixed in 10% neutral buffered formalin, stained with phosphotungstic acid and imaged by synchrotron micro-CT at beamline 8.3.2 of the Advanced Light Source (Lawrence Berkeley National Laboratory). All procedures were approved by the Penn State Institutional Animal Care and Use Committee (protocol PROTO202502937). The volume comprises 917 sagittal slices of 1052 x 6760 pixels at an isotropic voxel size of 0.5 micrometres, stored as 8-bit greyscale TIFF images (intensity 0-255). Annotations were drawn on every slice in 3D Slicer by expert annotators at Penn State College of Medicine and exported as TIFF label maps with the same dimensions and slice order as the raw volume. Each pixel carries one integer label: 0 = background, 1 = brain, 2 = heart, 3 = left eye, 4 = right eye. Where masks overlapped at boundaries, voxels were assigned to the heart. The data are highly class-imbalanced (background 97.55%, brain 1.36%, heart 0.09%, left eye 0.49%, right eye 0.50% of voxels), which makes them a useful benchmark for segmenting small structures in micro-CT. Contents: - raw/ : 917 greyscale micro-CT slices (TIFF) - labels/ : 917 five-class label maps (TIFF), matched to the raw slices by file name - split/ : lists of the 733 training and 184 held-out slice indices used in the paper (random slice-level split, seed 42)

Institutions

• Pennsylvania State University

  Pennsylvania, State College
• Penn State Milton S. Hershey Medical Center

  Pennsylvania, Hershey

Categories

Funders

• National Institutes of Health(NIH)

  Grant ID: R24OD037633

Licence