Adipocyte Cell Imaging Dataset
The dataset consists of images of the adipocytes (fat cells) taken using transmission light microscopy in the form of TIF files. There are three sets of images corresponding to three different magnification settings (20x, 40x, and 60x) of the microscope, with 50-100 images for each magnification setting. For each field of view there will be seven bright field images (transmission light images) for different values of focal planes, and three different fluorescence images corresponding to labeling of nuclei, lipocytes, and cell-matrix respectively. Each image is 2156 by 2556 pixels in size, using 16 bits to represent each pixel value.
Up to 300 pictures of stem-cell-derived human adipocyte cell cultures, using both brightfield and fluorescence imaging.
Collaboration between AstraZeneca and AI Sweden.
Real-world, highly challenging.
Use cases to date
1. Reconstructing the current bright field adipocyte images into deep learning enhanced new images that exhibit e.g., improved resolution, field-of-view, depth-of-field, etc., statistically matching the images that would be expected from higher-end imaging systems which are quite expensive.
2. The transmission light images do provide information about cellular organization, but they lack the clear contrast of fluorescence labeling which limits their use. However, fluorescence labeling can be labor-intensive and time-consuming. An AI solution can be used to transform the brightfield images into fluorescence images. This can be used in determining the location of 3D structures inside the cell, directly from transmitted light images.
→ Adipocyte Cell Imaging Dataset description
→ License Terms
→ Machine learning solution generated in AI Sweden and AstraZeneca challenge highlighted by Nature
→ How AI Sweden’s Data Factory enabled data scientists to leapfrog the drug production process
→ AI Sweden initiative contributes to accelerate the drug development process