In confocal microscopy, noise is an inevitable reality for some applications where it interferes with your research by obscuring fine details. Modern electronics coupled with unique detector designs significantly reduce the effects of readout noise. But for applications requiring short dwell times, noise from many sources gets amplified, making it necessary for you to use postprocessing algorithms to clean images. Many traditional and AI-based algorithms exist for denoising, but you need to be careful not to lose the truth during the denoising process.
This webinar provides a short introduction to signal-to-noise (S/N) measurement to evaluate a confocal laser scanning microscope. It then provides an overview of state-of-the-art denoising algorithms using classical and deep learning approaches, respectively. It also summarizes Noise2Void, which directly learns from noisy images without the need for respective clean images to train a deep learning model - making it ideal to denoise confocal images.