UIImage cv::Mat conversions with alpha channel - ios

Question

I'm using following codes for converting UIImage* and cv::Mat to each other:
- (cv::Mat)cvMatFromUIImage:(UIImage *)image
{
CGColorSpaceRef colorSpace = CGImageGetColorSpace(image.CGImage);
CGFloat cols = image.size.width;
CGFloat rows = image.size.height;
cv::Mat cvMat(rows, cols, CV_8UC4); // 8 bits per component, 4 channels (color channels + alpha)
CGContextRef contextRef = CGBitmapContextCreate(cvMat.data, // Pointer to data
cols, // Width of bitmap
rows, // Height of bitmap
8, // Bits per component
cvMat.step[0], // Bytes per row
colorSpace, // Colorspace
kCGImageAlphaNoneSkipLast |
kCGBitmapByteOrderDefault); // Bitmap info flags
CGContextDrawImage(contextRef, CGRectMake(0, 0, cols, rows), image.CGImage);
CGContextRelease(contextRef);
return cvMat;
}
and
-(UIImage *)UIImageFromCVMat:(cv::Mat)cvMat
{
NSData *data = [NSData dataWithBytes:cvMat.data length:cvMat.elemSize()*cvMat.total()];
CGColorSpaceRef colorSpace;
if (cvMat.elemSize() == 1) {
colorSpace = CGColorSpaceCreateDeviceGray();
} else {
colorSpace = CGColorSpaceCreateDeviceRGB();
}
CGDataProviderRef provider = CGDataProviderCreateWithCFData((__bridge CFDataRef)data);
// Creating CGImage from cv::Mat
CGImageRef imageRef = CGImageCreate(cvMat.cols, //width
cvMat.rows, //height
8, //bits per component
8 * cvMat.elemSize(), //bits per pixel
cvMat.step[0], //bytesPerRow
colorSpace, //colorspace
kCGImageAlphaNone|kCGBitmapByteOrderDefault,// bitmap info
provider, //CGDataProviderRef
NULL, //decode
false, //should interpolate
kCGRenderingIntentDefault //intent
);
// Getting UIImage from CGImage
UIImage *finalImage = [UIImage imageWithCGImage:imageRef];
CGImageRelease(imageRef);
CGDataProviderRelease(provider);
CGColorSpaceRelease(colorSpace);
return finalImage;
}
I took these from OpenCV Documentation. I use them as follows:
UIImage *img = [UIImage imageNamed:#"transparent.png"];
UIImage *img2 = [self UIImageFromCVMat:[self cvMatFromUIImage:img]];
However these functions loses the alpha channel information. I know it is because of the flags kCGImageAlphaNone and kCGImageAlphaNoneSkipLast, unfortunately I could't find a way not lose alpha information by changing these flags.
So, how do I convert these two types between each other without losing alpha information?
Here is the image that I use:

Answer 1

We should use these functions from opencv v2.4.6:
UIImage* MatToUIImage(const cv::Mat& image);
void UIImageToMat(const UIImage* image, cv::Mat& m, bool alphaExist = false);
And don't forget to include:
opencv2/imgcodecs/ios.h

Answer 2

You need to not pass kCGImageAlphaNoneSkipLast and instead pass (kCGBitmapByteOrder32Host | kCGImageAlphaPremultipliedFirst) to get premultiplied alpha in BGRA format. CoreGraphics only supports premultiplied alpha. But, you will need to check on how OpenCV represents alpha in pixels to determine how to tell OpenCV that the pixels are already premultiplied. The code I have used assumes straight alpha with OpenCV, so you will need to be careful of that.