Skip to content

Instantly share code, notes, and snippets.

Embed
What would you like to do?
A couple of swift functions for comparing two CGImage using CIImage in OS X
/**
@brief Returns true if images have same meta. Width, Height, bit depth.
@discussion Assumes images are non null.
*/
func doImagesHaveSameMeta(#image1:CGImage, #image2:CGImage) -> Bool {
if CGImageGetWidth(image1) != CGImageGetWidth(image2) {
return false
}
if CGImageGetHeight(image1) != CGImageGetHeight(image2) {
return false
}
if CGImageGetBitsPerComponent(image1) != CGImageGetBitsPerComponent(image2) {
return false
}
if CGImageGetBytesPerRow(image1) != CGImageGetBytesPerRow(image2) {
return false
}
if CGImageGetBitsPerPixel(image1) != CGImageGetBitsPerPixel(image2) {
return false
}
return true
}
/**
@brief Returns the maximum difference of pixel values in the image.
@discussion Assumes doImagesHaveSameMeta has already returned true on
the images passed into this function. OSX only as iOS doesn't have the
CIAreaMaximum filter.
*/
func compareImages(#image1:CGImage, #image2:CGImage) -> Int {
var diff = 0
// First create the CIImage representations of the CGImage.
let ciImage1 = CIImage(CGImage: image1)
let ciImage2 = CIImage(CGImage: image2)
// Create the difference blend mode filter and set its properties.
let diffFilter = CIFilter(name: "CIDifferenceBlendMode")
diffFilter.setDefaults()
diffFilter.setValue(ciImage1, forKey: kCIInputImageKey)
diffFilter.setValue(ciImage2, forKey: kCIInputBackgroundImageKey)
// Create the area max filter and set its properties.
let areaMaxFilter = CIFilter(name: "CIAreaMaximum")
areaMaxFilter.setDefaults()
areaMaxFilter.setValue(diffFilter.valueForKey(kCIOutputImageKey),
forKey: kCIInputImageKey)
let compareRect = CGRectMake(0.0, 0.0, CGFloat(CGImageGetWidth(image1)),
CGFloat(CGImageGetHeight(image1)))
let extents = CIVector(CGRect: compareRect)
areaMaxFilter.setValue(extents, forKey: kCIInputExtentKey)
// The filters have been setup, now set up the CGContext bitmap context the
// output is drawn to. Setup the context with our supplied buffer.
let alphaInfo = CGImageAlphaInfo.PremultipliedLast
let bitmapInfo = CGBitmapInfo(rawValue: alphaInfo.rawValue)
let colorSpace = CGColorSpaceCreateDeviceRGB()
var buf: [CUnsignedChar] = Array<CUnsignedChar>(count: 16, repeatedValue: 255)
let context = CGBitmapContextCreate(&buf, 1, 1, 8, 16, colorSpace, bitmapInfo)
// Now create the core image context CIContext from the bitmap context.
let ciContextOpts = [
kCIContextWorkingColorSpace : colorSpace,
kCIContextUseSoftwareRenderer : false
]
let ciContext = CIContext(CGContext: context, options: ciContextOpts)
// Get the output CIImage and draw that to the Core Image context.
let valueImage = areaMaxFilter.valueForKey(kCIOutputImageKey)! as CIImage
ciContext.drawImage(valueImage, inRect: CGRectMake(0,0,1,1),
fromRect: valueImage.extent())
// This will have modified the contents of the buffer used for the CGContext.
// Find the maximum value of the different color components. Remember that
// the CGContext was created with a Premultiplied last meaning that alpha
// is the fourth component with red, green and blue in the first three.
let maxVal = max(buf[0], max(buf[1], buf[2]))
diff = Int(maxVal)
return diff
}
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
You can’t perform that action at this time.