Fig. 1: System schematic diagram.

a, The MFIAS system comprises two components: selective illumination and focal modulation. The selective illumination component incorporates a DMD that is conjugated with the image plane and modulates the illumination patterns. The focal modulation component is a spinning disc with varying thicknesses of cover glass, which shifts the focal image to different depths. Before frame acquisition, a depth map is generated and converted into a series of illumination patterns displayed on the DMD. During each exposure period, the spinning disc completes one full rotation, while the DMD displays one pattern at each glass thickness. The resulting image is an integration of focal areas from different depths. TIR, total internal reflection prism; DM, dichroic mirror; OBJ, objective; LED, light-emitting diode; PE, photon emitter. b, Various techniques can be employed to generate illumination patterns. Image acquisition can be a stack for static or slow-varying surfaces or a single frame for fast-changing surfaces. The depth of each position can be determined using algorithms and/or refined through manual or semiautomatic detection algorithms. Red points in Step 2.3 are manually selected 3D coordinates. Finally, the depth map is generated, and the illumination pattern for each layer is calculated. CV, computer vision; DL, deep learning; LoG, Laplacian of Gaussian.