Resolution Reduction (Image Processing)


// Function for resolution reduction (useful for transforming/handling large images)
// Original code written by Peter Rehbein, Goethe-University, Frankfurt, Germany
Function Shrink(image, shrinkfactor)

Wave image
Variable shrinkfactor 

if(ShrinkFactor<=0)
	DoAlert /T="Alert" 0,"ShrinkFactor must be greater than 0!"
return -1
endif

Duplicate /O image adjusted_image
print "ShrinkFactor:"+num2str(shrinkfactor)
print "====="
Variable height = dimsize(adjusted_image,0)
Variable width = dimsize(adjusted_image,1)

print "Height (original):"+num2str(height)
print "Width (original):"+num2str(width)
print "====="
// Check whether image dimensions are dividable by shrink factor:

Variable check_height
Variable check_width

Variable new_height
Variable new_width

Variable u,v
Variable d,e

//Loop while height is not dividable by shrinkfactor and enlarge picture appropriately
//(New rows are filled with copied adjacent data)
check_height = mod(height,shrinkfactor)

if(check_height!=0)
print "Height not dividable by "+num2str(shrinkfactor)+"!"
	print "Starting correction!"
	print "====="
for(u=0;check_height!=0;u+=1)
	print "Actual Height: "+num2str(height)
	new_height = height+1
	Redimension/N=(new_height,-1) adjusted_image
		for(e=0;e<width;e+=1)
			//print "e: "+num2str(e)
			adjusted_image [new_height-1] [e] = adjusted_image [new_height-2] [e] 
		endfor
	height = new_height
	print "New Height: "+num2str(new_height)
	check_height = mod(height,shrinkfactor)
endfor
endif
// end check
	
//Loop while width is not dividable by shrinkfactor and enlarge picture appropriately
//(New columns are filled with copied adjacent data)
check_width = mod(width,shrinkfactor)
if(check_width!=0)
print "Width not dividable by "+num2str(shrinkfactor)+"!"
	print "Starting correction!"
	print "====="
for(v=0;check_width!=0;v+=1)
	print "Actual Width: "+num2str(width)
	new_width = width+1
	Redimension/N=(-1,new_width) adjusted_image
		for(d=0;d<height;d+=1)
			//print "d: "+num2str(d)
			adjusted_image [d] [new_width-1] = adjusted_image [d] [new_width-2] 
		endfor
	width = new_width
	print "New Width: "+num2str(new_width)
	check_width = mod(width,shrinkfactor)
endfor
endif
// end check

Variable sImage_Height = Height/ShrinkFactor
Variable sImage_Width = Width/ShrinkFactor
print "====="
print "SmallImage Height :"+num2str(sImage_Height)
print "SmallImage Width: "+num2str(sImage_Width)

Make /O /N=(sImage_Height, sImage_Width) sImage

Variable i,j,k,l

Variable avg

//print "====Report===="
for(i=0;i<sImage_Height;i+=1) //correct
	for(j=0;j<sImage_Width;j+=1) //correct		
	
		k=i*ShrinkFactor
		l=j*ShrinkFactor
				
		//print "===="
		//print "k:"+num2str(k)
		//print "l:"+num2str(l)
					
		Wave adjusted_image
		Duplicate/O/R=[k,k+ShrinkFactor-1][l,l+ShrinkFactor-1] adjusted_image,region // Extract region
		WaveStats /Q Region // Determine average pixel value
		avg = V_avg 
		//print "i: "+num2str(i)
		//print "j: "+num2str(j)
		//print "Average (region): "+num2str(avg)
		//print "i: "+num2str(i)
		//print "j: "+num2str(j)
		//print "===="
		sImage [i] [j] = avg // write average pixel value into smaller image	
	endfor
endfor

End
Not clear how this is different from using the built-in ImageInterpolate operation with the keyword Pixelate.
For example, in this part


Duplicate/O/R=[k,k+ShrinkFactor-1][l,l+ShrinkFactor-1] adjusted_image,region // Extract region
		WaveStats /Q Region // Determine average pixel value
		avg = V_avg 
		sImage [i] [j] = avg // write average pixel value into smaller image


you could exchange
V_avg
for
V_min
or
V_max
and write these values to the shrunken image instead :)
(This is actually what I wrote the snippet for)

Besides, it was fun to code ^^
One more suggestion:
it is usually a good idea to avoid memory allocations. Your code uses Duplicate followed by WaveStats. You could skip the Duplicate completely if you used ImageStats instead of WaveStats. Also, using either WaveStats or ImageStats you should be able to see some performance improvement with a /M=1 flag.

A.G.

Forum

Support

Gallery

Igor Pro 10

Learn More

Igor XOP Toolkit

Learn More

Igor NIDAQ Tools MX

Learn More