Hi,

I have just followed the video on integrating coreImage with SwiftUI. When simply loading in the image from the asset catalogue directly, the image is normally displayed. But when I run the image through UIImage, CIImage, apply a CIFilter of any kind, convert it to a CGFilter, and back to a SwiftUI Image, image is rotated 90% with the filter.

I followed the video 1:1 and tried to do a simple copy/paste of Paul's code in a new SwiftUI file, and the image is still roated.

Any thoughts on this?

import CoreImage
import CoreImage.CIFilterBuiltins
import SwiftUI

//CoreImage is about changing EXISTING images - not drawing new ones.
//An example image has been added to assets catalogue...

//Three types of Image types (CGImage, CIImage, UIImage) are required to work with if CoreImage is to be used to work with images

struct ContentView: View {
    @State private var image: Image?

    var body: some View {
        VStack {
            image?
                .resizable()
                .scaledToFit()
        }
        .onAppear(perform: loadImage)
    }

    func loadImage() {
        //Make the Swift Image image into a UIImage
        let inputImage = UIImage(resource: .IMG_1475)

        //Make the UIImage into a CIImage
        let beginImage = CIImage(image: inputImage)

        //Make a CI filter
        let context = CIContext()
        let currentFilter = CIFilter.crystallize()

        //Apply current filter to image
        currentFilter.inputImage = beginImage

        //make an amount for manipulating the filter
        let amount = 1.0

        //read the parameters that any given CIFilter supports
        let inputKeys = currentFilter.inputKeys

        //Apply that filter with the supported inputKeys for the selected filter onto the CIImage
        if inputKeys.contains(kCIInputIntensityKey) {
            currentFilter.setValue(amount, forKey: kCIInputIntensityKey) }
        if inputKeys.contains(kCIInputRadiusKey) { currentFilter.setValue(amount * 200, forKey: kCIInputRadiusKey) }
        if inputKeys.contains(kCIInputScaleKey) { currentFilter.setValue(amount * 10, forKey: kCIInputScaleKey) }

        //Converting the outputImage with its new filter to a Swift Image

        //Make a CI outputImage or bail
        guard let outputImage = currentFilter.outputImage else { return }

        //Make a CG image on the CI outputImage or bail
        guard let cgImage = context.createCGImage(outputImage, from: outputImage.extent) else { return }

        //Convert the CGImage into a UIImage
        let uiImage = UIImage(cgImage: cgImage)

        //Convert the UIImage into a the Swift Image we use in our view above
        image = Image(uiImage: uiImage)
    }
}

For anyone else running into this issue I added the following extension...

extension UIImage {
    func fixedOrientation() -> UIImage {
        if imageOrientation == .up {
            return self
        }

        UIGraphicsBeginImageContextWithOptions(size, false, scale)
        draw(in: CGRect(origin: .zero, size: size))
        let normalizedImage = UIGraphicsGetImageFromCurrentImageContext()!
        UIGraphicsEndImageContext()

        return normalizedImage
    }
}

I then added .fixedOrientation() to my inputImage & UIImage

guard let inputImage = UIImage(data: imageData)?.fixedOrientation() else { return }

&

let uiImage = UIImage(cgImage: cgImage).fixedOrientation()