View Peak Search

1. Consider a box around a local maximum with dimensions set by peaksizemm.
   
2. Let vmin be the minimum value in this box. (If peakmin exceeds vmin then vmin will be set to peakmin). The value of vmin will be subtracted from every pixel in the box.
   
3. Calculate:
   • vmax = maximum of the box
   • sum = ∑p (whole box)
   • n = number of pixels in the box
   • smallsum = ∑p (the nsmall central pixels)
   • nsmall = 9
     The number of central pixels, configured by peaksmallsize. A value of 1 leads to 9 central pixels, a value of 2 leads to 25 pixels.
   • ave = (sum-smallsum)/(n-nsmall)
   • smallave = (smallsum-vmax)/(nsmall-1)
   So ave is the average value without the central pixels and smallave is the average value of the central pixels without the single central pixel.
4. A peak is weak if
   • vmax < peakbg * ave
     So the maximum should be at least larger than peakbg times the average.
   • smallave < peakbg2 * ave
     So the average in the center should be stronger than than peakbg2 times the average.
5. A peak is too sharp if
   • vmax > peakbg3 * smallave
     So if the central pixel is larger than peakbg3 times the surrounding average.