I'm using VNDocumentViewController to scan some documents which is working fine. But not infrequently, the app becomes slow after dismissing the VNDocumentViewController. There is no VNDocumentViewController instance allocated, but according to the allocations gathered by instruments, there is a ICDocCamViewController still living and using between 200 and 300 MB. I guess that ICDocCamViewController is an internal component of VNDocumentViewController. Are there any known issues? Unfortunately, I do not see any way to free ICDocCamViewController. 281.37 MB 93.1% 104708 start_wqthread 281.37 MB 93.1% 104708 _pthread_wqthread 256.04 MB 84.7% 42747 _dispatch_workloop_worker_thread 256.03 MB 84.7% 42564 _dispatch_lane_invoke 256.03 MB 84.7% 42564 _dispatch_lane_serial_drain 249.91 MB 82.7% 22258 _dispatch_client_callout 237.01 MB 78.4% 8562 _dispatch_call_block_and_release 236.50 MB 78.3% 3836 __77-[ICDocCamViewController saveCapturedImage:metaData:rects:completionHandler:]_block_invoke_3 236.50 MB 78.3% 3836 -[ICDocCamViewController cropAndFilterImage:rects:filterType:] 236.49 MB 78.3% 3678 +[ICDocCamImageFilters filteredImage:orientation:imageFilterType:] 236.49 MB 78.3% 3678 +[ICDocCamImageFilters colorDocument:orientation:] 236.47 MB 78.3% 3476 -[CIContext(createCGImage) createCGImage:fromRect:] 236.47 MB 78.3% 3476 -[CIContext(_createCGImageInternal) _createCGImage:fromRect:format:premultiplied:colorSpace:deferred:renderCallback:] 236.40 MB 78.2% 2096 -[CIContext(CIRenderDestination) startTaskToRender:fromRect:toDestination:atPoint:error:] 236.40 MB 78.2% 2096 -[CIContext(CIRenderDestination) _startTaskToRender:toDestination:forPrepareRender:forClear:error:] 236.40 MB 78.2% 2096 CI::RenderToBitmap::render(CI::Image*, CI::Context*) const 236.40 MB 78.2% 2096 CI::image_get_bitmap(CI::Context*, CI::Image*, CGRect, CGColorSpace*, CI::Bitmap*, CI::RenderDestination const*) 236.33 MB 78.2% 716 CI::tile_node_graph(CI::Context*, CI::RenderDestination const*, char const*, CI::Node*, CGRect const&, CI::PixelFormat, CI::swizzle_info const&, CI::TileTask* (CI::ProgramNode*, CGRect) block_pointer) 236.33 MB 78.2% 716 CI::recursive_tile(CI::RenderTask*, CI::Context*, CI::RenderDestination const*, char const*, CI::Node*, CGRect const&, CI::PixelFormat, CI::swizzle_info const&, CI::TileTask* (CI::ProgramNode*, CGRect) block_pointer) 236.26 MB 78.2% 212 invocation function for block in CI::image_get_bitmap(CI::Context*, CI::Image*, CGRect, CGColorSpace*, CI::Bitmap*, CI::RenderDestination const*) 236.26 MB 78.2% 212 CI::Context::render(CI::ProgramNode*, CGRect const&) 236.26 MB 78.2% 212 CI::Context::recursive_render(CI::TileTask*, CI::roiKey const&, CI::Node*, bool) 236.26 MB 78.2% 211 CI::Context::recursive_render(CI::TileTask*, CI::roiKey const&, CI::Node*, bool) 236.24 MB 78.2% 196 CI::Context::recursive_render(CI::TileTask*, CI::roiKey const&, CI::Node*, bool) 236.23 MB 78.2% 181 CI::Context::recursive_render(CI::TileTask*, CI::roiKey const&, CI::Node*, bool) 236.21 MB 78.2% 166 CI::Context::recursive_render(CI::TileTask*, CI::roiKey const&, CI::Node*, bool) 236.19 MB 78.2% 151 CI::Context::recursive_render(CI::TileTask*, CI::roiKey const&, CI::Node*, bool) 17.02 KB 0.0% 14 CI::MetalTextureManager::create_intermediate(CI::IntermediateDescriptor const&, unsigned long long, CGRect const&, unsigned long, unsigned long, bool) 128 Bytes 0.0% 1 CI::MetalContext::render_intermediate_node(CI::TileTask*, CI::ProgramNode*, CGRect const&, CI::intermediate_t*, bool, void () block_pointer) 17.02 KB 0.0% 14 CI::MetalTextureManager::create_intermediate(CI::IntermediateDescriptor const&, unsigned long long, CGRect const&, unsigned long, unsigned long, bool) 128 Bytes 0.0% 1 CI::MetalContext::render_intermediate_node(CI::TileTask*, CI::ProgramNode*, CGRect const&, CI::intermediate_t*, bool, void () block_pointer) 17.02 KB 0.0% 14 CI::MetalTextureManager::create_intermediate(CI::IntermediateDescriptor const&, unsigned long long, CGRect const&, unsigned long, unsigned long, bool) 128 Bytes 0.0% 1 CI::MetalContext::render_intermediate_node(CI::TileTask*, CI::ProgramNode*, CGRect const&, CI::intermediate_t*, bool, void () block_pointer) 17.02 KB 0.0% 14 CI::MetalTextureManager::create_intermediate(CI::IntermediateDescriptor const&, unsigned long long, CGRect const&, unsigned long, unsigned long, bool) 128 Bytes 0.0% 1 CI::MetalContext::render_intermediate_node(CI::TileTask*, CI::ProgramNode*, CGRect const&, CI::intermediate_t*, bool, void () block_pointer) 128 Bytes 0.0% 1 CI::MetalContext::render_root_node(CI::TileTask*, CI::ProgramNode*, CGRect const&, void () block_pointer, void () block_pointer) 69.00 KB 0.0% 504 CI::gather_rois_for_program_graph(CI::Context*, char const*, CI::ProgramNode*, CGRect) 71.72 KB 0.0% 1380 CI::prepare_initial_graph(CI::Context*, char const*, CI::Image*, CI::RenderDestination const*, CGRect, CGColorSpace*, CI::PixelFormat, CI::swizzle_info, CI::Affine const&, bool, CI::TextureDescriptor*) 71.72 KB 0.0% 1380 CI::prepare_initial_graph(CI::Context*, char const*, CI::Image*, CI::RenderDestination const*, CGRect, CGColorSpace*, CI::PixelFormat, CI::swizzle_info, CI::Affine const&, bool, CI::TextureDescriptor*) 18.34 KB 0.0% 202 -[CIPaperWash outputImage]

I work on an iOS app that displays images that often contain text, and I'm adding support for ImageAnalysisInteraction as described in this WWDC 2022 session. I have gotten as far as making the interaction show up and being able to select text and get the system selection menu, and even add my own action to the menu via the buildMenuWithBuilder API. But what I really want to do with my custom action is get the selected text and do a custom lookup-like thing to check the text against other content in my app. So how do I get the selected text from an ImageAnalysisInteraction on a UIImageView? The docs show methods to check if there is selected text, but I want to know what the text is.

With the release of Xcode 13, a large section of my vision framework processing code became errors and cannot compile. All of these have became deprecated. This is my original code:  do {       // Perform VNDetectHumanHandPoseRequest       try handler.perform([handPoseRequest])       // Continue only when a hand was detected in the frame.       // Since we set the maximumHandCount property of the request to 1, there will be at most one observation.       guard let observation = handPoseRequest.results?.first else {         self.state = "no hand"         return       }       // Get points for thumb and index finger.       let thumbPoints = try observation.recognizedPoints(forGroupKey: .handLandmarkRegionKeyThumb)       let indexFingerPoints = try observation.recognizedPoints(forGroupKey: .handLandmarkRegionKeyIndexFinger)       let middleFingerPoints = try observation.recognizedPoints(forGroupKey: .handLandmarkRegionKeyMiddleFinger)       let ringFingerPoints = try observation.recognizedPoints(forGroupKey: .handLandmarkRegionKeyRingFinger)       let littleFingerPoints = try observation.recognizedPoints(forGroupKey: .handLandmarkRegionKeyLittleFinger)       let wristPoints = try observation.recognizedPoints(forGroupKey: .all)               // Look for tip points.       guard let thumbTipPoint = thumbPoints[.handLandmarkKeyThumbTIP],          let thumbIpPoint = thumbPoints[.handLandmarkKeyThumbIP],          let thumbMpPoint = thumbPoints[.handLandmarkKeyThumbMP],          let thumbCMCPoint = thumbPoints[.handLandmarkKeyThumbCMC] else {         self.state = "no tip"         return       }               guard let indexTipPoint = indexFingerPoints[.handLandmarkKeyIndexTIP],          let indexDipPoint = indexFingerPoints[.handLandmarkKeyIndexDIP],          let indexPipPoint = indexFingerPoints[.handLandmarkKeyIndexPIP],          let indexMcpPoint = indexFingerPoints[.handLandmarkKeyIndexMCP] else {         self.state = "no index"         return       }               guard let middleTipPoint = middleFingerPoints[.handLandmarkKeyMiddleTIP],          let middleDipPoint = middleFingerPoints[.handLandmarkKeyMiddleDIP],          let middlePipPoint = middleFingerPoints[.handLandmarkKeyMiddlePIP],          let middleMcpPoint = middleFingerPoints[.handLandmarkKeyMiddleMCP] else {         self.state = "no middle"         return       }               guard let ringTipPoint = ringFingerPoints[.handLandmarkKeyRingTIP],          let ringDipPoint = ringFingerPoints[.handLandmarkKeyRingDIP],          let ringPipPoint = ringFingerPoints[.handLandmarkKeyRingPIP],          let ringMcpPoint = ringFingerPoints[.handLandmarkKeyRingMCP] else {         self.state = "no ring"         return       }               guard let littleTipPoint = littleFingerPoints[.handLandmarkKeyLittleTIP],          let littleDipPoint = littleFingerPoints[.handLandmarkKeyLittleDIP],          let littlePipPoint = littleFingerPoints[.handLandmarkKeyLittlePIP],          let littleMcpPoint = littleFingerPoints[.handLandmarkKeyLittleMCP] else {         self.state = "no little"         return       }               guard let wristPoint = wristPoints[.handLandmarkKeyWrist] else {         self.state = "no wrist"         return       } ... } Now every line from thumbPoints onwards results in error, I have fixed the first part (not sure if it is correct or not as it cannot compile) to :         let thumbPoints = try observation.recognizedPoints(forGroupKey: VNHumanHandPoseObservation.JointsGroupName.thumb.rawValue)        let indexFingerPoints = try observation.recognizedPoints(forGroupKey: VNHumanHandPoseObservation.JointsGroupName.indexFinger.rawValue)        let middleFingerPoints = try observation.recognizedPoints(forGroupKey: VNHumanHandPoseObservation.JointsGroupName.middleFinger.rawValue)        let ringFingerPoints = try observation.recognizedPoints(forGroupKey: VNHumanHandPoseObservation.JointsGroupName.ringFinger.rawValue)        let littleFingerPoints = try observation.recognizedPoints(forGroupKey: VNHumanHandPoseObservation.JointsGroupName.littleFinger.rawValue)        let wristPoints = try observation.recognizedPoints(forGroupKey: VNHumanHandPoseObservation.JointsGroupName.littleFinger.rawValue) I tried many different things but just could not get the retrieving individual points to work. Can anyone help on fixing this?

Is there a framework that allows for classic image processing operations in real-time from incoming imagery from the front-facing cameras before they are displayed on the OLED screens? Things like spatial filtering, histogram equalization, and image warping. I saw the documentation for the Vision framework, but it seems to address high-level tasks, like object and recognition. Thank you!

How can I achieve full control over Vision Pro's display and effectively render a 2D graph plot on it? I would appreciate guidance on the necessary steps or code snippets. P.s. As per Apple documentation For a more immersive experience, an app can open a dedicated Full Space where only that app’s content will appear. This still does not fulfill the 'flat bounded 2D' requirement as the Spaces provide an unbounded 3D immersive view.

First of all this vision api is amazing. the OCR is very accurate. I've been looking to multiprocess using the vision API. I have about 2 million PDFs I want to OCR, and I want to run multiple threads/run parallel processing to OCR each. I tried pyobjc but it does not work so well. Any suggestions on tackling this problem?

Hi, Can VisionPro see outside and let us take a picture from code/iOS/Programming? Eg: like we used to have camera permission on iPad/iPhone to have access to device camera. Thank you

Trying to use VNGeneratePersonSegmentationRequest.. it seems to work but the output mask isn't at the same resolution as the source image.. so comping the result with the source produces a bad result. Not the full code, but hopefully enough to see what I'm doing. var imageRect = CGRect(x: 0, y: 0, width: image.size.width, height: image.size.height) let imageRef = image.cgImage(forProposedRect: &imageRect, context: nil, hints: nil)! let request = VNGeneratePersonSegmentationRequest() let handler = VNImageRequestHandler(cgImage: imageRef) do { try handler.perform([request]) guard let result = request.results?.first else { return } //Is this the right way to do this? let output = result.pixelBuffer //This ciImage alpha mask is a different resolution than the source image //So I don't know how to combine this with the source to cut out the foreground as they don't line up.. the res it's even the right aspect ratio. let ciImage = CIImage(cvPixelBuffer: output) ..... }

Why iOS 17's lift subject returns jagged edge and low resolution result image ? The output quality is totally different from iOS16. Does this just occur to beta version? or will it be the same in real iOS17 release version ?

My app uses the Vision framework to find images that are visually similar. Before WWDC23, this code worked predictably to request image observations and compare images: private func observe(cgImage: CGImage?) -> VNFeaturePrintObservation? { var returnVal: VNFeaturePrintObservation? = nil guard let cgImage = cgImage else {return returnVal} let imageRequestHandler = VNImageRequestHandler(cgImage: cgImage, options: [:]) let request = VNGenerateImageFeaturePrintRequest() request.usesCPUOnly = true do { try imageRequestHandler.perform([request]) returnVal = request.results?.first as? VNFeaturePrintObservation } catch { } return returnVal } func similarity(to compareAsset: Asset) -> Float { var dist = Float.infinity if let obs = self.observation, let compareObs = compareAsset.observation { try? obs.computeDistance(&dist, to: compareObs) } return dist } In the new frameworks, there is a new VNGenerateImageFeaturePrintRequestRevision value, and observations made with different request revisions can't be compared. If you try, you get an error: Error Domain=com.apple.vis Code=12 "The revision of the observations do not match" UserInfo={NSLocalizedDescription=The revision of the observations do not match}. The docs state that by explicitly setting the new VNGenerateImageFeaturePrintRequestRevision property on my requests, I can force the request to use a particular version. But I've updated the above code to do this, but explicitly setting the revision of my request doesn't work, and my app still gets tons of errors about mismatched request revisions. Here's the updated code: private func observe(cgImage: CGImage?) -> VNFeaturePrintObservation? { var returnVal: VNFeaturePrintObservation? = nil guard let cgImage = cgImage else {return returnVal} let imageRequestHandler = VNImageRequestHandler(cgImage: cgImage, options: [:]) let request = VNGenerateImageFeaturePrintRequest() if #available(iOS 17.0, *) { request.revision = VNGenerateImageFeaturePrintRequestRevision2 } else { request.revision = VNGenerateImageFeaturePrintRequestRevision1 } request.usesCPUOnly = true do { try imageRequestHandler.perform([request]) returnVal = request.results?.first as? VNFeaturePrintObservation } catch { print("\(type(of: self)) :: \(#function) :: error in observation request \(error)") } return returnVal } func similarity(to compareAsset: Asset) -> Float { var dist = Float.infinity if let obs = self.observation, let compareObs = compareAsset.observation { do { try obs.computeDistance(&dist, to: compareObs) } catch { print("\(type(of: self)) :: \(#function) :: error in simvalue \(error)") if (error as NSError).code == 12 { let revision = obs.requestRevision let compareRevision = compareObs.requestRevision print("\(type(of: self)) :: \(#function) :: simValue req mismatch \(revision) \(compareRevision)") } } } return dist } This breaks my app, and I can't figure out how to reliably force requests to the revision number I need. What am I doing wrong here? Will this behavior sort itself out as the SDK evolves? Thanks y'all

Lift subjects from images in your app, give a sample code and app

I got a crash when right click on an image in a web page loaded by WKWebView. The log and stack trace show that it might related to VisionKit. I think it may related to copy text on image feature. I don't know how to fix it. So I just want to disable VisionKit on WKWebView to check whether it fix the issue or not. Do you know how to do it? Stack trace: The log before crash: [com.apple.VisionKit.processing] Error processing request from MAD on result: Error Domain=NSOSStatusErrorDomain Code=-50 "paramErr: error in user parameter list" UserInfo={NSLocalizedDescription=Error allocating CVPixelBuffer} request: <VKCImageAnalyzerRequest: 0x6000025562b0> requestID: 22 madRequestID: (Not Set) cancelled: NO

Most of the APIs for this show iOS 16 support. Will a full subject lifting implementation require the device to be running iOS 17 or will it run on 16 too?

I'm using VisionKit framework for text recognition and detecting rectangles in the image. For that, I'm using VNRecognizeText & VNDetectRectangles features of the VisionKit. In macOS and iOS results, I found slight difference in the boundingBox coordinates of the text and the rectangles detected for the same image. Is this expected? Can we do anything to make the results identical? Also, on macOS, when I'm using same features of VisionKit from python (using pyobjc-framework-Vision package), there also i'm getting slightly different results.

When trying the code from this Session: https://developer.apple.com/videos/play/wwdc2023/10176/ Specifically: result.generateScaledMaskForImage(forInstances: result.allInstances, from: handler) It throws the error: -[VNInstanceMaskObservation generateScaledMaskForImageForInstances:fromRequestHandler:error:]: unrecognized selector sent to instance Any ideas?

Hi! Wondering if there's a way to add subject lifting to images in my SwiftUI app without relying on UIViewControllerRepresentable and Coordinators to adopt the ImageAnalysisInteraction protocol. Thank you!

Hi One point I would like to ask. How many specs do you need and which Mac would you recommend for Apple Vision Pro development? We use XCode, RealityKit, ARKit, Reality Composer Pro, Unity Editor which supports visionOS development, and MaterialX. If possible, what notebook and desktop models do you recommend? Best regards Sadao Tokuyama https://1planet.co.jp/

Hello guys, I am trying to run this sample project on my Ipad, I got a black screen the camera does not initialize. I tried updating the info.plist and asking for camera permission. I updated all the devices, did someone tried this demo? https://developer.apple.com/documentation/vision/detecting_animal_body_poses_with_vision

Hello there! Really excited about this year's additions to ImageAnalysisInteraction, especially the selectedText. Thank you! My question is if there's a way to have multiple interactions for a given image analysis. For example, instead of being able to make just one text selection, I would like to support multiple. In theory, I know multiple UIInteraction-conforming objects can be added to any view but I'm unsure if this is possible with ImageAnalysisInteraction. I'm willing to create a custom UIInteraction, if you see a potential way forward there (would love to hear any ideas!). Side question/feature request: UITextSelectionDisplayInteraction, which I assume ImageAnalysisInteraction might be using internally, allows handleViews customizations and others. I would be super useful if that was exposed through ImageAnalysisInteraction as well! Looking forward to hearing your ideas. TIA!

I have an AR app and soon will be making a VisionKit version. Can we use swift library to create a world space 3D UI with features like Zstack, Buttons, etc? Currently, I am using UI kit to do that.