In my iPhone app, I take a picture with the camera, then I want to resize it to 290*390 pixels. I was using this method to resize the image :
UIImage *newI
I've discovered that it's difficult to find an answer that you can use out-of-the box in your Swift 3 project. The main problem of other answers that they don't honor the alpha-channel of the image. Here is the technique that I'm using in my projects.
extension UIImage {
func scaledToFit(toSize newSize: CGSize) -> UIImage {
if (size.width < newSize.width && size.height < newSize.height) {
return copy() as! UIImage
}
let widthScale = newSize.width / size.width
let heightScale = newSize.height / size.height
let scaleFactor = widthScale < heightScale ? widthScale : heightScale
let scaledSize = CGSize(width: size.width * scaleFactor, height: size.height * scaleFactor)
return self.scaled(toSize: scaledSize, in: CGRect(x: 0.0, y: 0.0, width: scaledSize.width, height: scaledSize.height))
}
func scaled(toSize newSize: CGSize, in rect: CGRect) -> UIImage {
if UIScreen.main.scale == 2.0 {
UIGraphicsBeginImageContextWithOptions(newSize, !hasAlphaChannel, 2.0)
}
else {
UIGraphicsBeginImageContext(newSize)
}
draw(in: rect)
let newImage = UIGraphicsGetImageFromCurrentImageContext()
UIGraphicsEndImageContext()
return newImage ?? UIImage()
}
var hasAlphaChannel: Bool {
guard let alpha = cgImage?.alphaInfo else {
return false
}
return alpha == CGImageAlphaInfo.first ||
alpha == CGImageAlphaInfo.last ||
alpha == CGImageAlphaInfo.premultipliedFirst ||
alpha == CGImageAlphaInfo.premultipliedLast
}
}
Example of usage:
override func viewDidLoad() {
super.viewDidLoad()
let size = CGSize(width: 14.0, height: 14.0)
if let image = UIImage(named: "barbell")?.scaledToFit(toSize: size) {
let imageView = UIImageView(image: image)
imageView.center = CGPoint(x: 100, y: 100)
view.addSubview(imageView)
}
}
This code is a rewrite of Apple's extension with added support for images with and without alpha channel.
As a further reading I recommend checking this article for different image resizing techniques. Current approach offers decent performance, it operates high-level APIs and easy to understand. I recommend sticking to it unless you find that image resizing is a bottleneck in your performance.