Nearest neighbors deconvolution

Does anyone know of a plug-in that performs nearest neighbors deconvolution using an estimated rather than measured point spread function?

By estimated do you mean theoretical?? Or something else?

The PSF is generated using a separate algorithm, then is used as an input to most deconvolution algorithms. There is one cavaet for Nearest Neightbor it is a 2D algorithm so requires a 2D PSF.

You can generate theoretical PSFs using PSF Generator

Or you could experiment with using a Gaussian as the PSF.


Thanks, I’m sure I’m using poor terminology. I’ve been using the gaussian psf generator plug-in on images of small fluorescent beads with success, but i’ve been considering the output to be a measured psf. I was just wondering if there were any algorithms that estimated the psf based on the image stack to be de-convolved alone without needing the bead images. I’ve seen that option in some microscope control software referred to as nearest neighbors deconvolution, where the nearest neighbors are the stack slices above and below. thanks again

1 Like

Are your images from a widefield microscope or confocal?? Are they 3D or 2D?

If you know the NA of the lens, spacings of the image and EMW, you can use the PSF Generator plugin to calculate a theoretical PSF, without having a bead image.

If you are able to share an example image I can try processing it.

Thanks a lot for the advice. The images are 3d widefield, I managed to get the PSF data entered into the PSF generator and seem to be getting reasonable results. Many thanks for the offer to help.