GitLab

-fixed: incorrect line heights in z-levels selector component in optical scan, when in-/decreasing z-levels to many/low amounts

-fixed: when zooming in and clicking somewhere in MainWindow DetectionView sometimes does not show the correct image segment

-fixed: in OpticalScan mode the scenePyramid was called twice to update the scene after each successful scan of one stack of (scan)tiles
 * this may have lead to buffer overflows and null pointer exceptions (in qt5 and cv2) on Win10

-some additional logging
-fixed some index confusions (cell/pixel indexing in cv2 vs numpy) in ScenePyramid

This messes up the progress bar....
Also: more complete blockUI and unblockUI

A new parameter allows setting a maximum size of connected Components. Larger components are scaled down accordingly to avoid MemoryErrors.

A progress bar is implemented for visualizing the detection progress.
The progress bar is put in a dedicated class (in uielements.py) and it is now also used by opticalscan and ramanscanui.

-fixes tiles not being generated if full image is only slightly bigger than a multiple of tile size and scanned images not reaching into all tiles, so full image would not be fully covered by tiles

-original loadAndPastImage() function inserted again, renamed to legacyLoadAndPasteImage(), bc. it is still used by some components
-while scanning images will be saved in folder scanimages. these will now be deleted automatically
-fixed error in ScenePyramid::addSrcTile() when scanned images are spanning more than two tiles in either direction

Further performance improvements
Fixing error when cancelling detection (returning None only twice!)

Further performance improvements
Fixing error when cancelling detection (returning None only twice!)

-detectionview uses unscaled image

The segmentation now takes considerably less memory and can run on larger images.

(Optional) fuzzy clustering of seedpoints per component reduces over-segmentation.

Linearspace, logspce or completely customized

Plus: Fix in loading dataset