This video demonstrates how to use LandingLens to train a computer vision model for a multiclass classification project using an open-access retinopathy of prematurity dataset.
It covers the end-to-end process, including data preparation, model training with custom augmentations, performance evaluation, and deploying the model for API-based inference.
LandingLens achieves a F1 score of 92 on the test set as compared to 83 published by the original authors that provided the data.
Original paper with links to download the dataset: https://www.nature.com/articles/s41597-024-03362-5
Citation:
Zhao, X., Chen, S., Zhang, S. et al. A fundus image dataset for intelligent retinopathy of prematurity system. Sci Data 11, 543 (2024). https://www.nature.com/articles/s41597-024-03362-5
