Skip to content

Artificial labeling models

Joanna Pylvänäinen edited this page Oct 7, 2024 · 1 revision

Artificial labeling deep learning models


The models were trained using ZeroCostDL4Mic Pix2pix Notebooks.

Model Name Performance Purpose and Associated Figure Training Dataset Link
pix2pix_HUVEC_nuclei_cancer_cells_dataset SSIM = 0.755
lpips = 0.120
This model was used in Fig. 4 to artificially label nulcei from BF images with cancer and endothelial cells. Link
pix2pix_HUVEC_nuclei_immuno_cells_dataset SSIM = 0.756
lpips = 0.130
This model was used in Fig. 4 to artificially label nulcei from BF images with immuno and endothelial cells. Link
pix2pix_HUVEC_juctions_dataset SSIM = 0.270
lpips = 0.360
This model was used in Fig. 4 to artificially label cell-cell juctions from BF images with immuno or cancer and endothelial cells. Link