This technology includes a microscopy technique that produces super-resolution images from diffraction-limited images obtained from a line scanning confocal microscope. First, the operation of the confocal microscope is modified so that images with sparse line excitation are recorded. Second, these images are processed to increase resolution in one dimension. Third, by taking a series of such super-resolved images from a given sample type, a neural network may be trained to produce images with 1D super-resolution from new diffraction-limited images. Finally, by rotating a given diffraction-limited image through a series of angles, passing each of these images through the trained network, and combining the resulting 1D super-resolved images with joint deconvolution, images with isotropic super-resolution may be obtained.