Auto-focus methods and systems for digital imaging using multi-spectral trajectories

A system for acquiring focused images of a specimen on a slide, by determining optimal scanning trajectories, the system comprising:

• a digital imager that captures a relatively low magnification image of the slide to locate the specimen;
• and
• a digital imaging system that:
• forms a grid comprised of an arrangement of grid points;
• overlays at least part of the grid over a field of view that covers at least part of the specimen;
• for each grid point within the at least part of the grid:
• receives, from the digital imager, a set of Z-stack images that correspond to the grid point;
• computes an average gradient value for one or more image features of each image of the set of Z-stack images;
• selects an image from the set of Z-stack images having a greatest average gradient value;
• and
• processes the image having the greatest average gradient value to generate a focus value corresponding to the grid point;
• uses the focus value of each grid point within the at least part of the grid to form a resulting grid of three dimensional points;
• and
• based on the resulting grid, determines one or more three dimensional scanning trajectories.

A non-transitory computer-readable medium storing instructions which, when executed by one or more processors, cause a system for acquiring focused images of a specimen on a slide to perform operations comprising:

• instructing a digital imager to capture a relatively low magnification image of the slide, in order to locate the specimen;
• forming a grid comprised of an arrangement of grid points;
• overlaying at least part of the grid over a field of view that covers at least part of the specimen;
• for each grid point of the at least part of the grid:
• receiving, from the digital imager, a set of Z-stack images that correspond to the grid point;
• computing an average gradient value for one or more image features of each image of the set of Z-stack images;
• selecting an image from the set of Z-stack images having a greatest average gradient value;
• and
• processing the image having the greatest average gradient value to generate a focus value corresponding to the grid point
• using the focus value of each grid point within the at least part of the grid to form a resulting grid of three dimensional points;
• and
• based on the resulting grid, determining one or more three dimensional scanning trajectories.