Jonathan Ventura

The paper produces a four-year canopy cover time series for California using NAIP imagery and a U-Net model, showing no statistically significant trend in canopy change.

The paper presents a dataset for evaluating deep learning denoising methods on low-SNR fluorescence microscopy images and compares three models, finding the supervised transformer-based model performs best but requires more training time.

The paper provides high-quality fluorescence microscopy datasets for training deep neural networks to improve image quality and reduce photobleaching and phototoxicity.