Spot Detection

These are draft notes, a copy of notes given to user at Image Clinics. Feel free to add and remove material as well as reorganise the content

  • A good intro to data analysis, w. t-test etc

  • MATLAB

    • commands: mean, mode, median, std, ttest, ttest2, ....

  • In MATLAB if you ask to fit Gaussian to data (histogram data)

    • It really just gives you the mean and std

  • Spot detection

  • Learning image analysis

  • Good resources

  • Background subtraction

  • Suggestion

    • Set up appointment with ScopeM microscopy specialist and bring
      sample

    • Email Tobias, Joachim, Justine, (see ScopeM list of people)

    • Image analysis is much easier if the images are good => get good images!

      • junk in -> junk out

  • Workflow

    1. Illumination correction/background
      subtraction/background-correction/flat-field-correction (helps segmentation)

    2. Brightness and contrast adjustment (only for visualisation)

    3. Detect/segment spots - create binary mask

    4. Apply mask to original image, perhaps after illumination
      correction - but be careful if making measurements on an adjusted image

    5. Measure intensities through mask on the original, or something
      close to it - rule of thumb is to only measure on an image where
      artifacts have been corrected, i.e. no contrast enhancement but
      ok to adjust for uneven illumination

  • Overview

    • Ok to adjust image in all kind of ways, in order to detect
      particles

    • For measurements however, measure on the (artifact corrected) original