I'm trying to create a sky mask on pictures taken from my iPhone. I've seen in the documentation that CoreImage support semantic segmentation for Sky among other type for person (skin, hair etc...) For now, I didn't found the proper workflow to use it. First, I watched https://developer.apple.com/videos/play/wwdc2019/225/ I understood that images must be captured with the segmentation with this kind of code: photoSettings.enabledSemanticSegmentationMatteTypes = self.photoOutput.availableSemanticSegmentationMatteTypes photoSettings.embedsSemanticSegmentationMattesInPhoto = true I capture the image on my iPhone, save it as HEIC format then later, I try to load the matte like that : let skyMatte = CIImage(contentsOf: imageURL, options: [.auxiliarySemanticSegmentationSkyMatte: true]) Unfortunately, self.photoOutput.availableSemanticSegmentationMatteTypes always give me a list of types for person only and never a types Sky. Anyway, the AVSemanticSegmentationMatte.MatteType is just [Hair, Skin, Teeth, Glasses] ... No Sky !!! So, How am I supposed to use semanticSegmentationSkyMatteImage ?!? Is there any simple workaround ?

I have written two custom Core Image metal kernels which I'm using to produce a CIImage (by chaining several filters). I'm drawing the output image in a simple view and whatever I use (CIImage, NSImage, CGImageRef), the image appears corrupted on screen, like some sort of graphics corruption (I've tried on two different machines with different systems). However if I add a simple step to write the image to disk from the CIImage then read it from disk and draw it in that very same view, then all is fine and the image appears correctly. What could possibly be happening here?

When using the heif10Representation and writeHEIF10Representation APIs of CIContext, the resulting image doesn’t contain an alpha channel. When using the heifRepresentation and writeHEIFRepresentation APIs, the alpha channel is properly preserved, i.e., the resulting HEIC will contain a urn:mpeg:hevc:2015:auxid:1 auxiliary image. This image is missing when exporting as HEIF10. Is this a bug or is this intentional? If I understand the spec correctly, HEIF10 should be able to support alpha via auxiliary image (like HEIF8).

I am processing CVPixelBuffers received from camera using both Metal and CoreImage, and comparing the performance. The only processing that is done is taking a source pixel buffer and applying crop & affine transforms, and saving the result to another pixel buffer. What I do notice is CPU usage is as high a 50% when using CoreImage and only 20% when using Metal. The profiler shows most of the time spent is in CIContext render: let cropRect = AVMakeRect(aspectRatio: CGSize(width: dstWidth, height: dstHeight), insideRect: srcImage.extent) var dstImage = srcImage.cropped(to: cropRect) let translationTransform = CGAffineTransform(translationX: -cropRect.minX, y: -cropRect.minY) var transform = CGAffineTransform.identity transform = transform.concatenating(CGAffineTransform(translationX: -(dstImage.extent.origin.x + dstImage.extent.width/2), y: -(dstImage.extent.origin.y + dstImage.extent.height/2))) transform = transform.concatenating(translationTransform) transform = transform.concatenating(CGAffineTransform(translationX: (dstImage.extent.origin.x + dstImage.extent.width/2), y: (dstImage.extent.origin.y + dstImage.extent.height/2))) dstImage = dstImage.transformed(by: translationTransform) let scale = max(dstWidth/(dstImage.extent.width), CGFloat(dstHeight/dstImage.extent.height)) let scalingTransform = CGAffineTransform(scaleX: scale, y: scale) transform = CGAffineTransform.identity transform = transform.concatenating(scalingTransform) dstImage = dstImage.transformed(by: transform) if flipVertical { dstImage = dstImage.transformed(by: CGAffineTransform(scaleX: 1, y: -1)) dstImage = dstImage.transformed(by: CGAffineTransform(translationX: 0, y: dstImage.extent.size.height)) } if flipHorizontal { dstImage = dstImage.transformed(by: CGAffineTransform(scaleX: -1, y: 1)) dstImage = dstImage.transformed(by: CGAffineTransform(translationX: dstImage.extent.size.width, y: 0)) } var dstBounds = CGRect.zero dstBounds.size = dstImage.extent.size _ciContext.render(dstImage, to: dstPixelBuffer!, bounds: dstImage.extent, colorSpace: srcImage.colorSpace ) Here is how CIContext was created: _ciContext = CIContext(mtlDevice: MTLCreateSystemDefaultDevice()!, options: [CIContextOption.cacheIntermediates: false]) I want to know if I am doing anything wrong and what could be done to lower CPU usage in CoreImage?

Hello, I am a student and I am doing a search for my thesis on create ML and shape recognition and image processing, so for this subject I want to find the details of the steps used in create ML for this, such as the techniques used for pre-processing, and the methods of extracting characteristics, and the filters applied, ect...

I take a picture using the iPhone's camera. The taken resolution is 3024.0 x 4032. I then have to apply a watermark to this image. After a bunch of trial and error, the method I decided to use was taking a snapshot of a watermark UIView, and drawing that over the image, like so: // Create the watermarked photo. let result: UIImage=UIGraphicsImageRenderer(size: image.size).image(actions: { _ in   image.draw(in: .init(origin: .zero, size: image.size))   let watermark: Watermark = .init(     size: image.size,     scaleFactor: image.size.smallest / self.frame.size.smallest   )   watermark.drawHierarchy(in: .init(origin: .zero, size: image.size), afterScreenUpdates: true) }) Then with the final image — because the client wanted it to have a filename as well when viewed from within the Photos app and exported from it, and also with much trial and error — I save it to a file in a temporary directory. I then save it to the user's Photo library using that file. The difference as compared to saving the image directly vs saving it from the file is that when saved from the file the filename is used as the filename within the Photos app; and in the other case it's just a default photo name generated by Apple. The problem is that in the image saving code I'm getting the following error: [Metal] 9072 by 12096 iosurface is too large for GPU And when I view the saved photo it's basically just a completely black image. This problem only started when I changed the AVCaptureSession preset to .photo. Before then there was no errors. Now, the worst problem is that the app just completely crashes on drawing of the watermark view in the first place. When using .photo the resolution is significantly higher, so the image size is larger, so the watermark size has to be commensurately larger as well. iOS appears to be okay with the size of the watermark UIView. However, when I try to draw it over the image the app crashes with this message from Xcode: So there's that problem. But I figured that could be resolved by taking a more manual approach to the drawing of the watermark then using a UIView snapshot. So it's not the most pressing problem. What is, is that even after the drawing code is commented out, I still get the iosurface is too large error. Here's the code that saves the image to the file and then to the Photos library: extension UIImage {   /// Save us with the given name to the user's photo album.   /// - Parameters:   ///  - filename: The filename to be used for the saved photo. Behavior is undefined if the filename contain characters other than what is represented by this regular expression [A-Za-z0-9-_]. A decimal point for the file extension is permitted.   ///  - location: A GPS location to save with the photo.   fileprivate func save(_ filename: String, _ location: Optional<Coordinates>) throws {           // Create a path to a temporary directory. Adding filenames to the Photos app form of images is accomplished by first creating an image file on the file system, saving the photo using the URL to that file, and then deleting that file on the file system.     //   A documented way of adding filenames to photos saved to Photos was never found.     // Furthermore, we save everything to a `tmp` directory as if we just tried deleting individual photos after they were saved, and the deletion failed, it would be a little more tricky setting up logic to ensure that the undeleted files are eventually     // cleaned up. But by using a `tmp` directory, we can save all temporary photos to it, and delete the entire directory following each taken picture.     guard       let tmpUrl: URL=try {         guard let documentsDirUrl=NSSearchPathForDirectoriesInDomains(.documentDirectory, .userDomainMask, true).first else {           throw GeneralError("Failed to create URL to documents directory.")         }         let url: Optional<URL> = .init(string: documentsDirUrl + "/tmp/")         return url       }()     else {       throw GeneralError("Failed to create URL to temporary directory.")     }           // A path to the image file.     let filePath: String=try {               // Reduce the likelihood of photos taken in quick succession from overwriting each other.       let collisionResistantPath: String="\(tmpUrl.path(percentEncoded: false))\(UUID())/"               // Make sure all directories required by the path exist before trying to write to it.       try FileManager.default.createDirectory(atPath: collisionResistantPath, withIntermediateDirectories: true, attributes: nil)               // Done.       return collisionResistantPath + filename     }()     // Create `CFURL` analogue of file path.     guard let cfPath: CFURL=CFURLCreateWithFileSystemPath(nil, filePath as CFString, CFURLPathStyle.cfurlposixPathStyle, false) else {       throw GeneralError("Failed to create `CFURL` analogue of file path.")     }           // Create image destination object.     //     // You can change your exif type here.     //   This is a note from original author. Not quite exactly sure what they mean by it. Link in method documentation can be used to refer back to the original context.     guard let destination=CGImageDestinationCreateWithURL(cfPath, UTType.jpeg.identifier as CFString, 1, nil) else {       throw GeneralError("Failed to create `CGImageDestination` from file url.")     }           // Metadata properties.     let properties: CFDictionary={               // Place your metadata here.       // Keep in mind that metadata follows a standard. You can not use custom property names here.       let tiffProperties: Dictionary<String, Any>=[:]               return [         kCGImagePropertyExifDictionary as String: tiffProperties       ] as CFDictionary     }()           // Create image file.     guard let cgImage=self.cgImage else {       throw GeneralError("Failed to retrieve `CGImage` analogue of `UIImage`.")     }     CGImageDestinationAddImage(destination, cgImage, properties)     CGImageDestinationFinalize(destination)             // Save to the photo library.     PHPhotoLibrary.shared().performChanges({       guard let creationRequest: PHAssetChangeRequest = .creationRequestForAssetFromImage(atFileURL: URL(fileURLWithPath: filePath)) else {         return       }       // Add metadata to the photo.       creationRequest.creationDate = .init()       if let location=location {         creationRequest.location = .init(latitude: location.latitude, longitude: location.longitude)       }     }, completionHandler: { _, _ in       try? FileManager.default.removeItem(atPath: tmpUrl.absoluteString)     })   } } If anyone can provide some insight as to what's causing the iosurface is too large error and what can be done to resolve it, that'd be awesome.

Hello there 👋 I've noticed a different behavior between iOS 15 and iOS 16 using CIFilter and SpriteKit. Here is a sample code where I want to display a text and apply a blurry effect on the same text in the back of it. Here is the expected behavior (iOS 15): And the broken behavior on iOS 16: It looks like the text is rotated around the x-axis and way too deep. Here is the sample code: import UIKit import SpriteKit class ViewController: UIViewController {     var skView: SKView?     var scene: SKScene?     override func viewDidLoad() {         super.viewDidLoad()         skView = SKView(frame: view.frame)         scene = SKScene(size: skView?.bounds.size ?? .zero)         scene?.backgroundColor = UIColor.red         view.addSubview(skView!)         skView!.presentScene(scene)         let neonNode = SKNode()         let glowNode = SKEffectNode()         glowNode.shouldEnableEffects = true         glowNode.shouldRasterize = true         let blurFilter = CIFilter(name: "CIGaussianBlur")         blurFilter?.setValue(20, forKey: kCIInputRadiusKey)         glowNode.filter = blurFilter         glowNode.blendMode = .alpha         let labelNode = SKLabelNode(text: "MOJO")         labelNode.fontName = "HelveticaNeue-Medium"         labelNode.fontSize = 60         let labelNodeCopy = labelNode.copy() as! SKLabelNode         glowNode.addChild(labelNode)         neonNode.addChild(glowNode)         neonNode.addChild(labelNodeCopy)         neonNode.position = CGPoint(x: 200, y: 200)         scene?.addChild(neonNode) } }

I am using AVFoundation for live camera view. I can get my device from the current video input (of type AVCaptureDeviceInput) like: let device = videoInput.device The device's active format has a isPortraitEffectSupported. How can I set the Portrait Effect on and off in live camera view? I setup the camera like this: private var videoInput: AVCaptureDeviceInput! private let session = AVCaptureSession() private(set) var isSessionRunning = false private var renderingEnabled = true private let videoDataOutput = AVCaptureVideoDataOutput() private let photoOutput = AVCapturePhotoOutput() private(set) var cameraPosition: AVCaptureDevice.Position = .front func configureSession() { sessionQueue.async { [weak self] in guard let strongSelf = self else { return } if strongSelf.setupResult != .success { return } let defaultVideoDevice: AVCaptureDevice? = strongSelf.videoDeviceDiscoverySession.devices.first(where: {$0.position == strongSelf.cameraPosition}) guard let videoDevice = defaultVideoDevice else { print("Could not find any video device") strongSelf.setupResult = .configurationFailed return } do { strongSelf.videoInput = try AVCaptureDeviceInput(device: videoDevice) } catch { print("Could not create video device input: \(error)") strongSelf.setupResult = .configurationFailed return } strongSelf.session.beginConfiguration() strongSelf.session.sessionPreset = AVCaptureSession.Preset.photo // Add a video input. guard strongSelf.session.canAddInput(strongSelf.videoInput) else { print("Could not add video device input to the session") strongSelf.setupResult = .configurationFailed strongSelf.session.commitConfiguration() return } strongSelf.session.addInput(strongSelf.videoInput) // Add a video data output if strongSelf.session.canAddOutput(strongSelf.videoDataOutput) { strongSelf.session.addOutput(strongSelf.videoDataOutput) strongSelf.videoDataOutput.videoSettings = [kCVPixelBufferPixelFormatTypeKey as String: Int(kCVPixelFormatType_32BGRA)] strongSelf.videoDataOutput.setSampleBufferDelegate(self, queue: strongSelf.dataOutputQueue) } else { print("Could not add video data output to the session") strongSelf.setupResult = .configurationFailed strongSelf.session.commitConfiguration() return } // Add photo output if strongSelf.session.canAddOutput(strongSelf.photoOutput) { strongSelf.session.addOutput(strongSelf.photoOutput) strongSelf.photoOutput.isHighResolutionCaptureEnabled = true } else { print("Could not add photo output to the session") strongSelf.setupResult = .configurationFailed strongSelf.session.commitConfiguration() return } strongSelf.session.commitConfiguration() } } func prepareSession(completion: @escaping (SessionSetupResult) -> Void) { sessionQueue.async { [weak self] in guard let strongSelf = self else { return } switch strongSelf.setupResult { case .success: strongSelf.addObservers() if strongSelf.photoOutput.isDepthDataDeliverySupported { strongSelf.photoOutput.isDepthDataDeliveryEnabled = true } if let photoOrientation = AVCaptureVideoOrientation(interfaceOrientation: interfaceOrientation) { if let unwrappedPhotoOutputConnection = strongSelf.photoOutput.connection(with: .video) { unwrappedPhotoOutputConnection.videoOrientation = photoOrientation } } strongSelf.dataOutputQueue.async { strongSelf.renderingEnabled = true } strongSelf.session.startRunning() strongSelf.isSessionRunning = strongSelf.session.isRunning strongSelf.mainQueue.async { strongSelf.previewView.videoPreviewLayer.session = strongSelf.session } completion(strongSelf.setupResult) default: completion(strongSelf.setupResult) } } } Then to I set isPortraitEffectsMatteDeliveryEnabled like this: func setPortraitAffectActive(_ state: Bool) { sessionQueue.async { [weak self] in guard let strongSelf = self else { return } if strongSelf.photoOutput.isPortraitEffectsMatteDeliverySupported { strongSelf.photoOutput.isPortraitEffectsMatteDeliveryEnabled = state } } } However, I don't see any Portrait Effect in the live camera view! Any ideas why?

CVPixelBuffer.h defines kCVPixelFormatType_420YpCbCr8BiPlanarVideoRange = '420v', /* Bi-Planar Component Y'CbCr 8-bit 4:2:0, video-range (luma=[16,235] chroma=[16,240]). baseAddr points to a big-endian CVPlanarPixelBufferInfo_YCbCrBiPlanar struct */ kCVPixelFormatType_420YpCbCr10BiPlanarVideoRange = 'x420', /* 2 plane YCbCr10 4:2:0, each 10 bits in the MSBs of 16bits, video-range (luma=[64,940] chroma=[64,960]) */ But when I set above format camera output, and I find the output pixelbuffer's value is exceed the range.I can see [0 -255] for 420YpCbCr8BiPlanarVideoRange and [0,1023] for 420YpCbCr10BiPlanarVideoRange Is it a bug or something wrong of the output?If it is not how can I choose the correct matrix transfer the yuv data to rgb?

I am trying to use a CIColorKernel or CIBlendKernel with sampler arguments but the program crashes. Here is my shader code which compiles successfully. extern "C" float4 wipeLinear(coreimage::sampler t1, coreimage::sampler t2, float time) { float2 coord1 = t1.coord(); float2 coord2 = t2.coord(); float4 innerRect = t2.extent(); float minX = innerRect.x + time*innerRect.z; float minY = innerRect.y + time*innerRect.w; float cropWidth = (1 - time) * innerRect.w; float cropHeight = (1 - time) * innerRect.z; float4 s1 = t1.sample(coord1); float4 s2 = t2.sample(coord2); if ( coord1.x > minX && coord1.x < minX + cropWidth && coord1.y > minY && coord1.y <= minY + cropHeight) { return s1; } else { return s2; } } And it crashes on initialization. class CIWipeRenderer: CIFilter { var backgroundImage:CIImage? var foregroundImage:CIImage? var inputTime: Float = 0.0 static var kernel:CIColorKernel = { () -> CIColorKernel in let url = Bundle.main.url(forResource: "AppCIKernels", withExtension: "ci.metallib")! let data = try! Data(contentsOf: url) return try! CIColorKernel(functionName: "wipeLinear", fromMetalLibraryData: data) //Crashes here!!!! }() override var outputImage: CIImage? { guard let backgroundImage = backgroundImage else { return nil } guard let foregroundImage = foregroundImage else { return nil } return CIWipeRenderer.kernel.apply(extent: backgroundImage.extent, arguments: [backgroundImage, foregroundImage, inputTime]) } } It crashes in the try line with the following error: Fatal error: 'try!' expression unexpectedly raised an error: Foundation._GenericObjCError.nilError If I replace the kernel code with the following, it works like a charm: extern "C" float4 wipeLinear(coreimage::sample_t s1, coreimage::sample_t s2, float time) { return mix(s1, s2, time); }