Re: Fast Algorithm to Determine Average Brightness of an Image

[hh]

@BR

I adjusted your function a litte bit (and it works now here).
There is a new parameter for testing the upper/lower half
of the image (this param = empty tests the full image).
The mouseUp is my suggestion for a first testing. On base of
these two values you could develop a decision rule for a
dark/light setting.
[This may be better than simply averaging "sky" and "stones".
For example my first trial for a rule would be
if 0.62*upperBrightness + 0.38*lowerBrightness > 127 then
  set totalBrightnes to "light"
else set totalBrightness to "dark"]

on mouseUp
  set the paintcompression to RLE
  set topleft of img 1 to the bottomleft of me
  put the millisecs into m1
  put avgBrightness2(the long id of img 1,"upper") into b
  set backcolor of grc 1 to (b,b,b)
  put b into c
  put avgBrightness2(the long id of img 1,"lower") into b
  set backcolor of grc 2 to (b,b,b)
  put comma & b after c
  put c & ": " & (the millisecs - m1) & " ms" into fld 1
end mouseUp

## The adjusted function:

-- theImage = the long id of an img
-- thePart= "upper" or "lower" or empty (or anything else)
function avgBrightness2 theImage, thePart
  lock screen; lock messages
  put "tmp.42.42" into ii
  put the imageData of theImage into iData0
  put the width of img 1 into tOrigWidth
  if thePart is "upper" then
put (the height of img 1) div 2 into tOrigHeight
put byte 1 to 4*tOrigHeight*tOrigWidth of iData0 into iData0
  else if thePart is "lower" then
put (the height of img 1) div 2 into tOrigHeight
put byte -4*tOrigHeight*tOrigWidth to -1 of iData0 into iData0
  else
put the height of img 1 into tOrigHeight
  end if
  repeat while there is an img ii
delete img ii
  end repeat
  set the paintcompression to RLE -- important for speed
  create img ii
  set resizeQuality of img ii to "normal" -- "good", "best"
  set width of img ii to tOrigWidth   -- may be important*
  set height of img ii to tOrigHeight -- may be important*
  set imagedata of img ii to iData0
  put 8 into newBase -- choose newBase in range 1-64 (watch speed!)
  if tOrigWidth > tOrigHeight then
put newBase into newHeight
put trunc(newBase*tOrigWidth/tOrigHeight) into newWidth
  else
put newBase into newWidth
put trunc(newBase*tOrigHeight/tOrigWidth) into newHeight
  end if
  set width of img ii to newWidth
  set height of img ii to newHeight
  put the imagedata of img ii into iData
  delete img ii
  put 0 into vRed; put 0 into vGreen; put 0 into vBlue
  repeat with i=1 to newHeight
put (i-1)*4 into i0
repeat with j=1 to newWidth
  add byteToNum(byte i0+4*j-2 of iData) to vRed
  add byteToNum(byte i0+4*j-1 of iData) to vGreen
  add byteToNum(byte i0+4*j   of iData) to vBlue
end repeat
  end repeat
  -- return round((0.1*vRed+0.6*vGreen+0.3*vBlue)/newWidth/newHeight)
  return round(avg(vRed, vGreen, vRed)/newWidth/newHeight)
  unlock screen; unlock messages
end avgBrightness2

* Your experience of "deleting" or not img "tmp" was an effect of the
resizeQuality setting because you didn't set the tmp-image to the
original's size. This differs not this much from setting the size, but
it differs by a 'jumping' amount for different original's size and
different resizeQuality.

___
use-livecode mailing list
use-livecode@lists.runrev.com
Please visit this url to subscribe, unsubscribe and manage your subscription 
preferences:
http://lists.runrev.com/mailman/listinfo/use-livecode

Re: Fast Algorithm to Determine Average Brightness of an Image

[Sannyasin Brahmanathaswami]

@ HH

OK I will "bite" on your idea.  but I have never examined or processed the 
imagedata. Please scrutinize my code below.  But there is a major caveat with 
your original code (see below)

I think I got it correct. 

but there is a huge difference on the result between the 1 X 1 and the 12 X 9

I have a late afternoon shot of the sun setting over the ocean. your original 
function returns

148.981481 # which seems too low…

But if I did this correctly (I may not have)

I get 

192.33  # which seems a lot more correct visually because this is bright 
image.

average across 12x 9 = 108 px = 432 bytes of image data.

did I do something wrong here. (disclaimer… this could probably be 
optimized…but still very fast)


function avgBrightness theImage
   --lock screen; lock messages
   
   if there is no img ii then create img ii
   set resizeQuality of img ii to "best"
   
   put the width of img 1 into tOrigWidth
   put the height of img 1 into tOrigHeight
   
   set width of img ii to tOrigWidth
   set height of img ii to tOrigHeight
   
   set imagedata of img ii to the imagedata of img theImage
   
   put 12 into tNewWidth
   
   # Scale down proportinally
   set the width of img ii to tNewWidth
   put  round ( (tOrigHeight*tNewWidth)/tOrigWidth) into tNewHeight
   set height of img ii to tNewHeight
   
   --set width of img ii to 1
   --set height of img ii to 1
   
   put the imagedata of img ii into iData
   
   put tNewWidth * tNewHeight * 4 into tNoPixelBytes
   
   delete img ii
   
   # get beyond first byte otherwise mod 4 returns bogus value

   put (byteToNum(byte 2 of iData) & "," &\
   byteToNum(byte 3 of iData) & "," &\
   byteToNum(byte 4 of iData) & ",")   into tAllPixelColorValues
   
   repeat with x = 5 to tNoPixelBytes
  if x mod 4 = 1 then next repeat # skip this one, all zeros here.
  put byteToNum(byte x of iData) &"," after tAllPixelColorValues
   end repeat

   put empty into item -1 of tAllPixelColorValues # remove last empty itm
   
  return avg (tAllPixelColorValues)
  unlock screen; unlock messages
   end avgBrightness
BR


But if you use your original code  and do not delete img ii, a different value 
is returned than if you uncomment that line.. if you delete img ii you get a 
different value.



function avgBrightness1 theImage
   --lock screen; lock messages
   
  if there is no img ii then create img ii
  set resizeQuality of img ii to "best"
  
  put the width of img 1 into tOrigWidth
  put the height of img 1 into tOrigHeight
  
  set imagedata of img ii to the imagedata of img theImage
  set width of img ii to 1
  set height of img ii to 1
 put the imagedata of img ii into iData
 --delete img ii
  return avg (byteToNum(byte 2 of iData), \
byteToNum(byte 3 of iData), \
byteToNum(byte 4 of iData))
  --unlock screen; unlock messages
end avgBrightness1
 

On 9/20/16, 1:31 PM, "use-livecode on behalf of hh" 
 wrote:

There is still one option that uses a weighted grayLevel-mean
(I use these weights in imageJIT). This would reflect more
than simple averaging that one wants a dark/light decision.

return (0.1*byteToNum(byte 2 of iData) \
  + 0.6*byteToNum(byte 3 of iData) \
  + 0.3*byteToNum(byte 4 of iData))

Also, what will be still fast enough, you could think about 
scaling down _proportional_ to 3x2 pixels (or 12x8 pixels)
and then applying the above weighted mean on these few pixels.

p.s. I checked: The one-pixel color is by LC built from the
arithmetic mean of the color channels, nearly as fast as an external!



___
use-livecode mailing list
use-livecode@lists.runrev.com
Please visit this url to subscribe, unsubscribe and manage your subscription 
preferences:
http://lists.runrev.com/mailman/listinfo/use-livecode

Re: Fast Algorithm to Determine Average Brightness of an Image

[hh]

There is still one option that uses a weighted grayLevel-mean
(I use these weights in imageJIT). This would reflect more
than simple averaging that one wants a dark/light decision.

return (0.1*byteToNum(byte 2 of iData) \
  + 0.6*byteToNum(byte 3 of iData) \
  + 0.3*byteToNum(byte 4 of iData))

Also, what will be still fast enough, you could think about 
scaling down _proportional_ to 3x2 pixels (or 12x8 pixels)
and then applying the above weighted mean on these few pixels.

p.s. I checked: The one-pixel color is by LC built from the
arithmetic mean of the color channels, nearly as fast as an external!

___
use-livecode mailing list
use-livecode@lists.runrev.com
Please visit this url to subscribe, unsubscribe and manage your subscription 
preferences:
http://lists.runrev.com/mailman/listinfo/use-livecode

Re: Fast Algorithm to Determine Average Brightness of an Image

[Sannyasin Brahmanathaswami]

@ hh

That works!

fascinating. Though what is bright not bright, is, as you say, "not sharp" 

Images that return a value > 100 still seem bright enough to warrant darker 
type… 

and of course you may have e.g. very dark stone mountains (taking up 1/3 of the 
composition space at the bottom half of the image) with a very bright, blue 
clear sky… 

So "average brightness" has only so much utility in terms of choosing a color 
brightness for an overlay.

But, this is still very useful function.  Thank you!
 

On 9/20/16, 10:14 AM, "use-livecode on behalf of hh" 
 wrote:

You could try the following function. It is "dirty" (the definition
of 'brightness' is also unsharp!), but it is very fast.

local ii="tmp-1.414214"

-- uses the average color intensity of the pixel
-- resulting from scaling down to a 1x1 image
function avgBrightness theImage
  lock screen; lock messages
  if there is no img ii then create img ii
  set resizeQuality of img ii to "best"
  set width of img ii to the width of img 1
  set height of img ii to the height of img 1
  set imagedata of img ii to the imagedata of theImage
  set width of img ii to 1
  set height of img ii to 1
  put the imagedata of img ii into iData
  delete img ii
  return avg (byteToNum(byte 2 of iData), \
byteToNum(byte 3 of iData), \
byteToNum(byte 4 of iData))
  unlock screen; unlock messages
end avgBrightnes

___
use-livecode mailing list
use-livecode@lists.runrev.com
Please visit this url to subscribe, unsubscribe and manage your subscription 
preferences:
http://lists.runrev.com/mailman/listinfo/use-livecode

Re: Fast Algorithm to Determine Average Brightness of an Image

[hh]

> So is there a fast, efficient way to analyze a 1200 X 800 px
> image to get this average brightness value.

You could try the following function. It is "dirty" (the definition
of 'brightness' is also unsharp!), but it is very fast.

local ii="tmp-1.414214"

-- uses the average color intensity of the pixel
-- resulting from scaling down to a 1x1 image
function avgBrightness theImage
  lock screen; lock messages
  if there is no img ii then create img ii
  set resizeQuality of img ii to "best"
  set width of img ii to the width of img 1
  set height of img ii to the height of img 1
  set imagedata of img ii to the imagedata of theImage
  set width of img ii to 1
  set height of img ii to 1
  put the imagedata of img ii into iData
  delete img ii
  return avg (byteToNum(byte 2 of iData), \
byteToNum(byte 3 of iData), \
byteToNum(byte 4 of iData))
  unlock screen; unlock messages
end avgBrightness


___
use-livecode mailing list
use-livecode@lists.runrev.com
Please visit this url to subscribe, unsubscribe and manage your subscription 
preferences:
http://lists.runrev.com/mailman/listinfo/use-livecode

Re: Fast Algorithm to Determine Average Brightness of an Image

[Rick Harrison]

Hi Sannyasin,

Based on my previous experiences trying to process
images with LC pixel by pixel, I would have to say no.
LC is too slow for this kind of processing.

You would probably have to use a routine written in
some other language like Objective C, or C++, etc.

If anyone else has had a different experience with this type
of image processing in LC please chime in!

Thanks,

Rick

> On Sep 20, 2016, at 3:15 PM, Sannyasin Brahmanathaswami  
> wrote:
> 
> Use Case:
> 
> I want to place puzzle tiles on top of an image. We take a field, put a grc 
> behind it, take a snapshot, crop each word off at -1000, -1000 and then 
> scramble the tiles to make a puzzle… all this works really well. So I decided 
> to dress it up a bit more: we dynamically choose a color for the foreground 
> of the field in RGB range of 0-135 for all three colors → This gets us dark 
> type… then add 100 to all the colors and this gets us a color for the bkgnd 
> grc which is the same hue, but brighter (well, that depends on your preferred 
> brand HSV, HSL, HSB LAB color theory, it's not that simple… but, close enough 
> for this context)  tiles look great.. color is randomized for each puzzle 
> with the caveat of an occasional "ugly" (in relation to the nature photo in 
> the background)  but still… cool..
> 
> OK I thought hmmm we could also use light type with dark background and I got 
> this working… too much fun..
> 
> Switch sUserColorRange 
> case "darkType"
> set the itemdelimiter to "," # just to be safe
> put randomInRange(0,135) into tRGB
> put comma & randomInRange(0,135) after tRGB
> put comma & randomInRange(0,135) after tRGB
> set the backgroundcolor of grc "startHere" to tRGB
> set the foregroundcolor of fld _Quote to tRGB
> # same hue for background, but light
> repeat for each item tHue in tRGB
> put tHue+ (115) & comma after tTileBkgndRGB
> end repeat
> delete char -1 of tTileBkgndRGB
> set the backgroundColor of grc quoteBkgnd to tTileBkgndRGB
> break
> case "lightType"
> set the itemdelimiter to "," # just to be safe
> put randomInRange(136,255) into tRGB
> put comma & randomInRange(136,255) after tRGB
> put comma & randomInRange(136,255) after tRGB
> set the backgroundcolor of grc "startHere" to tRGB  ## oops need to fix this… 
> too light, disappears
> set the foregroundcolor of fld _Quote to tRGB
> # same hue for background, but light
> repeat for each item tHue in tRGB
> put tHue- (115) & comma after tTileBkgndRGB
> end repeat
> delete char -1 of tTileBkgndRGB
> set the backgroundColor of grc quoteBkgnd to tTileBkgndRGB
> end Switch
> 
> The incoming nature photo that is the backdrop for the puzzle is a random 
> selection. The above is predicated on a user choice/save-as-setting, but a 
> preferred algorithm would be:
> 
> Determine average overall brightness of the image. If below 130 on a scale of 
> 1-255, then use light colored type on dark tile… if >130 then use dark type 
> on light tiles
> 
> So is there a fast, efficient way to analyze a 1200 X 800 px image to get 
> this average brightness value.
> 
> 
> 
> ___
> use-livecode mailing list
> use-livecode@lists.runrev.com
> Please visit this url to subscribe, unsubscribe and manage your subscription 
> preferences:
> http://lists.runrev.com/mailman/listinfo/use-livecode


___
use-livecode mailing list
use-livecode@lists.runrev.com
Please visit this url to subscribe, unsubscribe and manage your subscription 
preferences:
http://lists.runrev.com/mailman/listinfo/use-livecode

Fast Algorithm to Determine Average Brightness of an Image

[Sannyasin Brahmanathaswami]

Use Case:

I want to place puzzle tiles on top of an image. We take a field, put a grc 
behind it, take a snapshot, crop each word off at -1000, -1000 and then 
scramble the tiles to make a puzzle… all this works really well. So I decided 
to dress it up a bit more: we dynamically choose a color for the foreground of 
the field in RGB range of 0-135 for all three colors → This gets us dark type… 
then add 100 to all the colors and this gets us a color for the bkgnd grc which 
is the same hue, but brighter (well, that depends on your preferred brand HSV, 
HSL, HSB LAB color theory, it's not that simple… but, close enough for this 
context)  tiles look great.. color is randomized for each puzzle with the 
caveat of an occasional "ugly" (in relation to the nature photo in the 
background)  but still… cool..

OK I thought hmmm we could also use light type with dark background and I got 
this working… too much fun..

Switch sUserColorRange 
case "darkType"
set the itemdelimiter to "," # just to be safe
put randomInRange(0,135) into tRGB
put comma & randomInRange(0,135) after tRGB
put comma & randomInRange(0,135) after tRGB
set the backgroundcolor of grc "startHere" to tRGB
set the foregroundcolor of fld _Quote to tRGB
# same hue for background, but light
repeat for each item tHue in tRGB
put tHue+ (115) & comma after tTileBkgndRGB
end repeat
delete char -1 of tTileBkgndRGB
set the backgroundColor of grc quoteBkgnd to tTileBkgndRGB
break
case "lightType"
set the itemdelimiter to "," # just to be safe
put randomInRange(136,255) into tRGB
put comma & randomInRange(136,255) after tRGB
put comma & randomInRange(136,255) after tRGB
set the backgroundcolor of grc "startHere" to tRGB  ## oops need to fix this… 
too light, disappears
set the foregroundcolor of fld _Quote to tRGB
# same hue for background, but light
repeat for each item tHue in tRGB
put tHue- (115) & comma after tTileBkgndRGB
end repeat
delete char -1 of tTileBkgndRGB
set the backgroundColor of grc quoteBkgnd to tTileBkgndRGB
end Switch

The incoming nature photo that is the backdrop for the puzzle is a random 
selection. The above is predicated on a user choice/save-as-setting, but a 
preferred algorithm would be:

Determine average overall brightness of the image. If below 130 on a scale of 
1-255, then use light colored type on dark tile… if >130 then use dark type on 
light tiles

So is there a fast, efficient way to analyze a 1200 X 800 px image to get this 
average brightness value.



___
use-livecode mailing list
use-livecode@lists.runrev.com
Please visit this url to subscribe, unsubscribe and manage your subscription 
preferences:
http://lists.runrev.com/mailman/listinfo/use-livecode