Phase-based computational adaptive optics enables artifact-free super-resolution microscopy.

Adaptive optics has revolutionized biological microscopy by improving resolution and signal-to-noise ratio, yet its reliance on complex hardware and phototoxic wavefront sensing limits broader adoption. Here, we introduce ∅CAO, a computational phase-based adaptive optics technique that corrects optical aberrations in three-dimensional fluorescence microscopy without requiring specialized optics or training datasets. By leveraging phase transfer functions in the frequency domain, ∅CAO enables robust post-acquisition correction across diverse imaging modalities, including wide-field and structured illumination microscopy. Our method achieves substantial improvements in image fidelity, supports subregional aberration correction, and maintains performance under noisy conditions. Demonstrated on a range of biological specimens, including Caenorhabditis elegans and plant tissues, ∅CAO offers a scalable and accessible solution for high-resolution biological imaging, facilitating the broad deployment of adaptive optics approaches across the life sciences.