Generative Modeling with Conditional Autoencoders: Building an Integrated Cell
This work addresses the challenge of probabilistic interpretation and generalization in cell imaging for biological research, representing an incremental advancement in generative modeling for microscopy data.
The authors tackled the problem of modeling variation in cell and nuclear morphology and subcellular structure localization from microscopy images using a conditional generative model, resulting in the ability to produce photo-realistic cell images and predict unobserved structures.
We present a conditional generative model to learn variation in cell and nuclear morphology and the location of subcellular structures from microscopy images. Our model generalizes to a wide range of subcellular localization and allows for a probabilistic interpretation of cell and nuclear morphology and structure localization from fluorescence images. We demonstrate the effectiveness of our approach by producing photo-realistic cell images using our generative model. The conditional nature of the model provides the ability to predict the localization of unobserved structures given cell and nuclear morphology.