Dataset containing mouse brain sections with dopaminergic neuronal soma manually annotated
Description
This cell annotation dataset consists of 114 8-bit RGB images with associated 16-bit cell label images. The label images consist of a total of 18193 manually labeled tyrosine hydroxylase positive cells. For the label images, each cell area is assigned a pixel value in the range of 1 to n, where n is the total number of labeled cells in a label image. Since the label images are in 16-bit, it may appear black and pre-processing (for example, adjusting brightness/contrast) is needed to see the cells. For all the images, the image resolution is 0.46 microns per pixel.
Files
Steps to reproduce
The data-set used in this study was obtained by manually labeling TH positive DA neuronal soma in 2D histology digital images. This is an internal GNE data-set. The digital images were obtained from different animal studies where mouse brains were sectioned at 35 micron thickness and stained with TH and either Haematoxylin or Nissl as a background tissue stain. Each animal study consisted of multiple animals and all the animals in one study were stained at one time-frame. This allowed us to take into account the minor tissue processing and staining associated variability that commonly occurs in histology studies. The sections were then imaged using a whole slide scanner microscope, Nanozoomer system (Hamamatsu Corp, San Jose,3CA) at 20x resolution (0.46 microns/pixel). Whole coronal brain section images containing the Substantia Nigra were exported from the digital scans at 20x resolution and were used to annotate the TH positive DA neuronal soma and train the model. The ground truth (GT) for this study was labelled and quality controlled by biologists who specialize in mouse brain anatomy and PD research. We have evaluated human annotator bias in one of our previous neuroanatomy segmentation studies that have been published (https://doi.org/10.1016/j.neuri.2023.100131). A neuroanatomy expert further QCd the annotations and made the necessary changes to reflect the cell annotation. (Adapted from https://openreview.net/forum?id=izFnURFG3f) This cell annotation dataset consists of 114 8-bit RGB images with associated 16-bit cell label images. The label images consist of a total of 18193 manually labeled tyrosine hydroxylase positive cells. For the label images, each cell area is assigned a pixel value in the range of 1 to n, where n is the total number of labeled cells in a label image. Since the label images are in 16-bit, it may appear black and pre-processing (for example, adjusting brightness/contrast) is needed to see the cells. For all the images, the image resolution is 0.46 microns per pixel.