Contribution per pixel to a region of interest's average value

Geonostalgia

When drawing a region of interest, the region's area adjusts dynamically in terms of the size (measured in µm²) depending on where the lines sit amongst the pixels. This leads me to wonder - does this affect how much that pixel's value contributes to a region's average value for a given element? Is a ROI's average element value pixel-area weighted? For instance, if you only nick the corner of one pixel, does it only contribute a tiny amount to an ROI's element value average, or since it is technically "selected" for the region is that pixel's value wholesale added to the ROI's average value without any modification?
Thanks y'all!

Bence

Hi @Geonostalgia

Because of the way images are drawn with CellSpace, you could artificially create very pixel-dense (i.e. high resolution) images and calculate stats from those with artificially high n values. To avoid this, Joe came up with the brilliant idea of going back to the mass spec measurements, rather than doing stats on actual pixels. The result is that if you have an ROI that encompasses, for example, x values from 100 to 200, and y values from 50 to 150, iolite will look for all mass spec measurements with x and y values in those ranges. It then calculates the stats from those mass spec measurements and the underlying pixel grid of the image is ignored.

Regarding whether an ROI includes a measurement at the very edge, yes, if it is in range, it will be included in the stats.

I hope that helps.

Kind regards,
Bence

Geonostalgia

That makes sense - thanks for the reply! Just to be clear though, this means that the µm² area is the only thing that really relies on/corresponds to the pixel "grid" on the images directly (as opposed to it just being a way of reference certain mass spec time slices)?

Bence

Geonostalgia this means that the µm2 area is the only thing that really relies on/corresponds to the pixel "grid" on the images directly

It depends on how the image is constructed and what type of ROI it is, but the resolution of the pixel grid also doesn't factor into the area calculation. Importantly, if you're not using the CellSpace approach, the values entered into the 'Shape and Orientation' part of the Imaging Setup will determine the pixel size, which is used in the area calculations.

If it's a CellSpace image, and you have a polygonal ROI (i.e. drawn using the mouse), the area of the polygon(s) can be calculated directly and this is the area value you see in the ROI table.

If it is a disjoint ROI (such as one created from criteria), the pixel size is used (calculated using either the laser log details or the data entered in the 'Size and Orientation' fields). The number of pixels that comprise the disjoint ROI is multiplied by the pixel size to calculate the area.

I hope that helps. Please let us know if you have any other questions.

Kind regards,
Bence