It is structured as a wizard that guides the user through the denoising process in a sequence of steps. a In ImageJ, the user loads an image or image stack for denoising and starts the DenoisEM plugin. The UI wizard appears, and the computational backend for parallel processing on the CPU/GPU (Quasar) is initialized. b In the next step, the user chooses any of the open images for denoising and selects a ROI on which denoising will be previewed. c Next, the main panel in the plugin appears. At the top it shows side-by-side the noisy original as well as the denoised version of the selected ROI. In the bottom left corner, the user can select one of eight denoising algorithms. The bottom right has controls for specifying the algorithm parameters. Typically, if the algorithm or its parameters are changed, the denoised ROI at the top is updated virtually instantaneously. This allows the user to easily assess the effect of algorithms and parameter settings. d After optimal settings are chosen, the user is shown a short summary and (for image stacks) can select the image slices that need to be denoised. During denoising, the user is shown progress feedback. e When denoising is finished, a new image or image stack is created and displayed. The original image (stack) is left untouched.