Curvature and surfaces
Fit smooth surfaces to height maps and point clouds, then report mean and Gaussian curvature with uncertainty. Rings, ridges and necks get numbers.
Quantitative microscopy image analysis
Michael Imaging supplies the computational side of microscopy research: surface reconstruction, curvature, hole statistics, spatial patterns and the pipelines that make each result repeatable. Your group keeps the microscope.
Areas where careful mathematics and code turn microscopy images into measurements.
Fit smooth surfaces to height maps and point clouds, then report mean and Gaussian curvature with uncertainty. Rings, ridges and necks get numbers.
Detect, size and count openings in membranes and films. Diameter distributions, area fractions and spacing between neighbours.
Ripley’s K, the pair correlation g(r) and point-pattern tests for clustering, regularity and colocalization.
Python code with a parameter file, pinned software versions and a git commit behind every result. Any figure can be regenerated.
From ongoing research partnerships to a single well-defined task.
Michael Imaging joins a research group as the quantitative partner for microscopy data. We contribute computationally and leave the experiments to your team. Where a project calls for it, AI models can be part of the work too.
A narrow task with a clear deliverable, agreed before work starts. Examples:
Conversations about quantitative problems that need new mathematics. Some become joint papers, grant-funded work or a longer partnership.
A measurement is only useful if someone else can repeat it. Projects are set up so they can.
Pixel size, smoothing scale and fit options are recorded in the project, so nobody has to remember them.
A figure or table can be traced to the exact code version that produced it.
Curvatures, diameters and counts come with error estimates and the assumptions behind them.
Scripts and intermediate results are delivered, so your group can rerun or extend the analysis.
# recorded with every run pixel_size_um: 0.065 smooth_sigma_px: 2.0 surface_fit: thin_plate_spline hole_min_area_um2: 0.05 curvature: [mean, gaussian] bootstrap_n: 500 git_commit: 4e1c9a7 python: 3.12