Invention:
This invention is a data augmentation method to enhance a biomedical imaging dataset used for deep learning applications. The approach involves using animal models of a disease combined with an unpaired domain translation framework to convert animal data to the human domain. The method is novel and valuable in that it increases biological diversity in the dataset, overcoming a fundamental barrier in developing deep learning models for biomedical imaging applications.
Background:
One key barrier to applying deep learning to biological datasets is data scarcity. This fundamental challenge is mainly relevant in research involving genetically constrained organisms, organelles, specialized cell types, and biological cycles and pathways. This data augmentation method adds biological diversity to a dataset to enable further research in deep learning models for biomedical imaging applications.
Applications:
- Biomedical imaging
- Artificial intelligence
- Research and development
Advantages:
- Converts animal data to the human domain
- Increases biological diversity in dataset
- Novel method
