Application Notes

Nikon NIS-Elements Denoise.ai Software: utilizing deep learning to denoise confocal data

January 2020

Noise is a fundamental component of confocal images, a result of discreet digital sampling of continuously emitting photons from samples. The contribution of noise to image quality (signal-to-noise ratio) increases as the signal decreases as a square-root function. Using a trained neural network, we use artificial intelligence to remove the shot noise component from confocal image data, allowing an increase in image quality and the ability to acquire dimmer samples at faster rates. NIS-Elements software’s Denoise.ai deploys this trained network for live or post-acquisition processing.