We have developed a stripe noise reduction protocol, called DeStripe, for single biomolecule images obtained by AFM that are contaminated by heavy and thin bands of vertical noise. To detect and restore noisy pixel, the program adopts a "divide and conquer" strategy by dividing the Fourier spectrum of the image into a central region and a peripheral region. The method is also applicable to other noisy images with high density bands such as those acquired by scanning electron microscopy. The noise reduction effect introduced by DeStripe allows for better visualization of objects without introducing additional artifacts into the restored image. Theoretical concept and development performed by Wendy Chen.
Example of image processing :
Raw image__________________Denoised image____________Noise image