Training Data for XAcGAN Model

The Cross-Attention conditional generative adversarial networks (XAcGAN) model is a proposed system to provide an alternative to fluorescent staining protocols. Focussing on the morphology of the nuclei during the process of apoptosis, the XAcGAN model generates images illustrating the apoptopic nuclei derived from only bright field images. The uploaded images were the ground truth used to train the model. They represent CHO-K1 cells with the nuclei, mitochondria and actin filaments stained with hoechst, mitospy and phalloidin respectively. These cells were induced to undergo apoptosis via exposure to 1 µM of staurosporine. The image sets consist of z-stacks, 7.2 µm in length and consisting of 24 slices spaced 0.3 µm apart.

Complete download (zip, 1.6 GiB)

Alternative title Training data for Cross-Attention conditional generative adversarial networks model
Creator(s) Ruixiong Wang, Daniel Butt
Contributor(s) Paul Verkade, Alin Achim, Stephen Cross
Publication date 09 Nov 2022
Language eng
Publisher University of Bristol
Licence Non-Commercial Government Licence for public sector information
DOI 10.5523/bris.2w8trsx55b0k22qez7uycg9sf4
Citation Ruixiong Wang, Daniel Butt (2022): Training Data for XAcGAN Model. https://doi.org/10.5523/bris.2w8trsx55b0k22qez7uycg9sf4
Total size 1.6 GiB

Data Resources