Stable and efficient autofocus based on fuzzy edge detection in fluorescence images

Automatic focusing in fluorescence microscopy is crucial for high-resolution imaging but faces challenges due to noise, low contrast, and computational inefficiency. Existing autofocusing algorithms suffer from unstable peaks and interference in defocused images. This study introduces a fluorescence image fuzzy edge sharpness evaluation (FESE) algorithm based on a 5 × 5 pixel convolutional template for focus prediction. The dataset acquired using an Olympus microscope demonstrates the high sensitivity, stability, and efficiency of our technique, yielding distinct unimodal peaks. The method significantly improves focus detection accuracy and efficiency, making a unique contribution to advanced fluorescence microscopy.