diff --git a/Graphics/Filters/GD.hs b/Graphics/Filters/GD.hs new file mode 100644 --- /dev/null +++ b/Graphics/Filters/GD.hs @@ -0,0 +1,193 @@ +module Graphics.Filters.GD +( + -- * Filters + brightness + , colorize + , contrast + , gaussianBlur + , grayscale + , edgeDetect + , emboss + , meanRemoval + , negative + , smoothing + -- * Pixel transformation functions + , pixelTransform + , convolute +) where + +import Graphics.GD +import Graphics.Filters.Util +import Control.Monad (mapM_,foldM) + +{- | + Performs the supplied transform function on every pixel of the image. + The transform function should take as it's argument a quadruple of Ints (RGBA) and returns a new RGBA quadruple + which is will be the new RGBA values of the the pixel. + + For example, the 'colorize' filter's inner workings are implemented with this as the transform function: + + @ + (\\(r,g,b,a) -> let + nr = clamp 0 255 (r+ar) + ng = clamp 0 255 (g+ag) + nb = clamp 0 255 (b+ab) + na = clamp 0 127 (a+aa) + in (nr,ng,nb,na)) + @ +-} +pixelTransform :: Image + -> ( RGBA -> RGBA ) -- ^ Transform function to be performed on each pixel + -> IO () +pixelTransform img fx = do + (width,height) <- imageSize img + mapM_ (\y -> + mapM_ (\x -> do + curr <- getPixel (x,y) img + let + (r,g,b,a) = toRGBA curr + (nr,ng,nb,na) = fx (r,g,b,a) + setPixel (x,y) (rgba nr ng nb na) img + ) [0..(width-1)] + ) [0..(height-1)] + return () + +{- | + Performs the convolution matrix on each pixel of the original image. + After the matrix has been applied, the resulting RGBA value has each of it's elements divided by the Divisor argument + and then the Offset argument is added to each element + + For example, the 'emboss' filter is implemented with the following convolution: + + > emboss img = convolute img [[1.5,0.0,0.0],[0.0,0.0,0.0],[0.0,0.0,-1.5]] 1 127 +-} +convolute :: Image + -> [[Float]] -- ^ Convolution matrix + -> Float -- ^ Divisor + -> Float -- ^ Offset + -> IO () +convolute img matrix fdiv offset = do + (width,height) <- imageSize img + imgCpy <- copyImage img + mapM_ (\y -> + mapM_ (\x -> convoluteImage img imgCpy matrix fdiv offset x y) [0..(width-1)] + ) [0..(height-1)] + return () + +convoluteImage :: Image -> Image -> [[Float]] -> Float -> Float -> Int -> Int -> IO () +convoluteImage img imgCpy matrix fdiv offset x y = do + (nr,ng,nb,na) <- foldM (\(or,og,ob,oa) j -> do + let yy = min (max (y-(1+j)) 0) (max (y-1) 0) + (pr,pg,pb,pa) <- foldM (\(ar,ag,ab,aa) k -> do + let xx = min (max (x-(1+k)) 0) (max (x-1) 0) + curr <- getPixel (xx,yy) imgCpy + let (r,g,b,a) = toRGBA curr + return (ar + fromIntegral r * ((matrix!!j)!!k) + ,ag + fromIntegral g * ((matrix!!j)!!k) + ,ab + fromIntegral b * ((matrix!!j)!!k) + ,fromIntegral a) + ) (or,og,ob,oa) [0.. (length (matrix!!j) - 1)] + return (pr,pg,pb,pa) + ) ((0.0,0.0,0.0,0.0) :: (Float,Float,Float,Float)) [0.. (length matrix - 1)] + let + new_r = clamp 0 255 . truncate $ (nr/fdiv)+offset + new_g = clamp 0 255 . truncate $ (ng/fdiv)+offset + new_b = clamp 0 255 . truncate $ (nb/fdiv)+offset + setPixel (x,y) (rgba new_r new_g new_b (truncate na)) img + +{- | + Applies the supplied color transformation to the image. + The range of the passed RGB values are -255 to +255, and the range of the A value is -127 to +127. +-} +colorize :: Image -> RGBA -> IO () +colorize img (ar,ag,ab,aa) = + if or [ar > 255, ar < (-255), ag > 255, ag < (-255), ab > 255, ab < (-255), aa > 127, aa < (-127)] then + error "Argument out of bounds. Colorize expects the color argument's RGB components to be between -255 and +255\ + \ and it's A component between -127 and +127" + else + pixelTransform img (\(r,g,b,a) -> let + nr = clamp 0 255 (r+ar) + ng = clamp 0 255 (g+ag) + nb = clamp 0 255 (b+ab) + na = clamp 0 127 (a+aa) + in (nr,ng,nb,na)) + +{- | + Inverts the image's color. +-} +negative :: Image -> IO () +negative img = + pixelTransform img (\(r,g,b,a) -> let + nr = abs 255-r + ng = abs 255-g + nb = abs 255-b + in (nr,ng,nb,a)) + +{- | + Converts an image to grayscale +-} +grayscale :: Image -> IO () +grayscale img = + pixelTransform img (\(r,g,b,a) -> let + newcol = truncate $ 0.299 * fromIntegral r + 0.587 * fromIntegral g + 0.114 * fromIntegral b + in (newcol, newcol, newcol, a)) + +{- | + Applies the supplied contrast adjustment to the image. + The range of the contrast argument is -100 to +100, with + -100 being maximum contrast, and +100 being minimum contrast +-} +contrast :: Image -> Int -> IO () +contrast img contVal = + if or [contVal > 100, contVal < (-100)] then + error "Argument out of bounds. Contrast expects the contrast argument to be between -100 and +100" + else + pixelTransform img (\(r,g,b,a) -> let + contr = (**2) $ (100.0 - fromIntegral contVal )/100.0 + (ur,ug,ub,_) = normalizeRGBA (r,g,b,a) + nr = clamp 0 255 $ (*) 255 . (+) 0.5 . (*) contr $ ur-0.5 + ng = clamp 0 255 $ (*) 255 . (+) 0.5 . (*) contr $ ug-0.5 + nb = clamp 0 255 $ (*) 255 . (+) 0.5 . (*) contr $ ub-0.5 + in (truncate nr, truncate ng, truncate nb, a)) + +{- | + Applies the supplied brightness adjustment to the image. + The range of the brightness argument is -255 to +255 +-} +brightness :: Image -> Int -> IO () +brightness img brightVal = + if or [brightVal > 255, brightVal < (-255)] then + error "Argument out of bounds. Brightness expects the brightness argument to be between -255 and +255" + else + colorize img (brightVal,brightVal,brightVal,0) + +{- | + Applies Gaussian blur to the image +-} +gaussianBlur :: Image -> IO () +gaussianBlur img = convolute img [[1.0,2.0,1.0],[2.0,4.0,2.0],[1.0,2.0,1.0]] 16 0 + +{- | + Applies an Emboss effect to the image +-} +emboss :: Image -> IO () +emboss img = convolute img [[1.5,0.0,0.0],[0.0,0.0,0.0],[0.0,0.0,-1.5]] 1 127 + +{- | + Applies Edge Detection to the image +-} +edgeDetect :: Image -> IO () +edgeDetect img = convolute img [[-1.0,0.0,-1.0],[0.0,4.0,0.0],[-1.0,0.0,-1.0]] 1 127 + +{- | + Applies a Mean Removal effect to the image +-} +meanRemoval :: Image -> IO () +meanRemoval img = convolute img [[-1.0,-1.0,-1.0],[-1.0,9.0,-1.0],[-1.0,-1.0,-1.0]] 1 0 + +{- | + Applies weighted Smoothing to the image. The smoothing amount is technically unbounded, + but larger values produce a less noticeable result +-} +smoothing :: Image -> Float -> IO () +smoothing img weighting= convolute img [[1.0,1.0,1.0],[1.0,weighting,1.0],[1.0,1.0,1.0]] (weighting+8.0) 0 diff --git a/Graphics/Filters/Util.hs b/Graphics/Filters/Util.hs new file mode 100644 --- /dev/null +++ b/Graphics/Filters/Util.hs @@ -0,0 +1,19 @@ +module Graphics.Filters.Util where + +type RGBA = (Int,Int,Int,Int) + +-- | Utility function for clamping a value between a minimum and maximum value +clamp :: (Ord a, Num a) => a -- ^ Minimum + -> a -- ^ Maximum + -> a -- ^ Value to clamp + -> a +clamp minm maxm num + | num < minm = minm + | num > maxm = maxm + | otherwise = num + +-- | Function for converting RGBA from an Int 0-255 range to a Float 0.0-1.0 range +normalizeRGBA :: RGBA -> (Float,Float,Float,Float) +normalizeRGBA (r,g,b,a) = (fromIntegral r/255,fromIntegral g/255,fromIntegral b/255,fromIntegral a/127) + + diff --git a/LICENSE b/LICENSE new file mode 100644 --- /dev/null +++ b/LICENSE @@ -0,0 +1,30 @@ +Copyright (c)2012, Tim Channell + +All rights reserved. + +Redistribution and use in source and binary forms, with or without +modification, are permitted provided that the following conditions are met: + + * Redistributions of source code must retain the above copyright + notice, this list of conditions and the following disclaimer. + + * Redistributions in binary form must reproduce the above + copyright notice, this list of conditions and the following + disclaimer in the documentation and/or other materials provided + with the distribution. + + * Neither the name of Tim Channell nor the names of other + contributors may be used to endorse or promote products derived + from this software without specific prior written permission. + +THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS +"AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT +LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR +A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT +OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, +SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT +LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, +DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY +THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT +(INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE +OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. diff --git a/Setup.hs b/Setup.hs new file mode 100644 --- /dev/null +++ b/Setup.hs @@ -0,0 +1,2 @@ +import Distribution.Simple +main = defaultMain diff --git a/imagefilters.cabal b/imagefilters.cabal new file mode 100644 --- /dev/null +++ b/imagefilters.cabal @@ -0,0 +1,24 @@ +Name: imagefilters +Version: 0.1 +Synopsis: Image Filters (contrast, brightness, gaussian blur, etc) +Description: + This package brings image filters comparable to PHP's imagefilter functions. In the current release, + these filters may only be used with the GD Haskell extension library, but in the future it may support + other image management libraries, such as the new JuicyPixels library. + . + imagefilters also supplies functionality for creating custom filters by using the 'pixelTransform' function + for per-pixel transformations, or the 'convolute' function for the application of convolution matrices + +Homepage: https://github.com/tchannel/imagefilters +License: BSD3 +License-file: LICENSE +Author: Tim Channell +Maintainer: tim.channell@gmail.com +Category: Graphics +Build-type: Simple +Cabal-version: >=1.2 + +Library + Exposed-modules: Graphics.Filters.Util, Graphics.Filters.GD + Build-depends: base >= 4 && < 5, gd >= 3000 +