Page History

...

In some experiments not all sensors contain useful data. It might be more efficient to select Region of Interest (ROI) on sensors, where data need to be processed. To support this feature a tuple (or list) of windows is passed as a constructor parameter. Each window is presented by the tuple of 5 parameters (segnum, rowmin, rowmax, colmin, colmax), where segnum is a sensor index in the n-d array, other parameters constrain window 2-d matrix rows and columns. Several windows can be defined for the same sensor using the same segnum. For 2-d arrays segnumparameter is not used, but still needs to be presented in the window tuple by any integer number. To increase algorithm efficiency only pixels in windows are processed. If windows=None, all sensors will be processed.

The array of windows can be converted in 3-d or 2-d array of mask using method pyimgalgos.GlobalUtils.mask_from_windows.

Mask

Alternatively ROI can be defined by the mask of good/bad (1/0) pixels. For 2-d image mask can easily be defined in user's code. In case of ≥3-d arrays the Mask Editor helps to produce ROI mask. Entire procedure includes

...