IMAGE PROCESSING METHOD FOR ENHANCING RESOLUTION IN CONFORTENT SCANNING FLUORESCENCE MICROSCOPES USING DEEP LEARNING TECHNIQUES TO RECONSTRUCT IMAGES OF ANNELIC POINT BLURRING FUNCTIONS AND SUBTRACTION OF TWO IMAGES

Abstract

This invention proposes an image processing method to enhance the resolution of confocal scanning fluorescence microscopy. The method uses deep learning techniques to reconstruct the image of the speckled blur function of a traditional fluorescence microscopy into an image corresponding to the annular blur function, where the annular blur function is generated by a helical phase mask with a variation from 0 to 2π. Next, the resulting subtracted image is obtained by subtracting the image of the speckled blur function from the image generated by the deep learning model, thereby improving the resolution compared to the original image of the confocal scanning fluorescence microscopy. The proposed method allows for improved image resolution without changing the operating principle of the microscopy system or the image acquisition time.

Images