I am doing a livestream where i need to send video from camera and an overlay from UIView It started working But Fps go down after some seconds.
If i send CMSampleBuffer directly FPS is Ok But if convert CMSampleImage and perform editing on it FPS decreases after some seconds.
i am attaching the code below where i get CMSampleBuffer from captureOutput(_ output: AVCaptureOutput, didOutput sampleBuffer: CMSampleBuffer, from connection: AVCaptureConnection)
I am also attaching FPS graph from facebook livestream sdk.
func captureOutput(_ output: AVCaptureOutput, didOutput sampleBuffer: CMSampleBuffer, from connection: AVCaptureConnection) {
let pts:CMTime = CMSampleBufferGetPresentationTimeStamp(sampleBuffer) as CMTime
let newPts = CMTimeMakeWithSeconds(CMTimeGetSeconds(pts) + 5, preferredTimescale: pts.timescale);
let image = self.imageFromSampleBuffer(sampleBuffer: sampleBuffer)
let drawableRect = self.window.size.width > 800 ? CGRect(x: 73, y: 0, width: self.window.size.width - 147, height: self.window.size.height) : CGRect(x: 0, y: 0, width: self.window.size.width, height: self.window.size.height)
let webViewImage: UIImage = self.webview.screenShotWithoutDrawHierarchy(drawableRect: drawableRect) ?? UIImage()
let compositeImage = self.composite(image: image!, overlay: webViewImage, drawableRect: drawableRect)
if #available(iOS 13, *){
let newSampleBuffer = compositeImage?.createCMSampleBuffer(presentationTimeStamp: newPts, duration:CMTime.invalid, decodeTimeStamp: sampleBuffer.decodeTimeStamp)
self.rtmpStream.appendSampleBuffer(newSampleBuffer!, withType: .video)
}
}
func composite(image:UIImage, overlay:(UIImage), scaleOverlay: Bool = false, drawableRect:CGRect)->UIImage?{
UIGraphicsBeginImageContext(drawableRect.size)
image.draw(in: drawableRect)
UIColor(red: 0, green: 0, blue: 0, transparency: 0)?.setFill()
overlay.draw(in: drawableRect)
let newimage = UIGraphicsGetImageFromCurrentImageContext()
UIGraphicsEndImageContext()
return newimage
}
private func imageFromSampleBuffer(sampleBuffer: CMSampleBuffer) -> UIImage? {
guard let imageBuffer = CMSampleBufferGetImageBuffer(sampleBuffer) else { return nil }
let ciImage = CIImage(cvPixelBuffer: imageBuffer)
return self.convert(cmage:ciImage)
}
// Convert CIImage to UIImage
func convert(cmage: CIImage) -> UIImage {
let context = CIContext(options: nil)
let cgImage = context.createCGImage(cmage, from: cmage.extent)!
let image = UIImage(cgImage: cgImage)
return image
}
// clean up AVCapture
func stopCamera(){
session.stopRunning()
}
}
extension UIImage {
var cvPixelBuffer: CVPixelBuffer? {
let attrs = [
String(kCVPixelBufferCGImageCompatibilityKey): kCFBooleanTrue,
String(kCVPixelBufferCGBitmapContextCompatibilityKey): kCFBooleanTrue
] as [String: Any]
var buffer: CVPixelBuffer?
let window = UIApplication.shared.keyWindow!
let drawableRect = window.size.width > 800 ? CGRect(x: 73, y: 0, width: window.size.width - 147, height: window.size.height) : CGRect(x: 0, y: 0, width: window.size.width, height: window.size.height)
let status = CVPixelBufferCreate(kCFAllocatorDefault, Int(drawableRect.size.width), Int(drawableRect.size.height), kCVPixelFormatType_32ARGB, attrs as CFDictionary, &buffer)
guard status == kCVReturnSuccess else {
return nil
}
CVPixelBufferLockBaseAddress(buffer!, CVPixelBufferLockFlags(rawValue: 0))
let pixelData = CVPixelBufferGetBaseAddress(buffer!)
let rgbColorSpace = CGColorSpaceCreateDeviceRGB()
let context = CGContext(data: pixelData, width: Int(window.size.width), height: Int(window.size.height), bitsPerComponent: 8, bytesPerRow: CVPixelBufferGetBytesPerRow(buffer!), space: rgbColorSpace, bitmapInfo: CGImageAlphaInfo.noneSkipFirst.rawValue)
context?.translateBy(x: 0, y: window.size.height)
context?.scaleBy(x: 1.0, y: -1.0)
let newRect = CGRect(x: 0, y: 0, width: window.size.width, height: window.size.height)
UIGraphicsPushContext(context!)
UIColor.clear.setFill()
UIRectFill(newRect)
self.draw(in: newRect)
UIGraphicsPopContext()
CVPixelBufferUnlockBaseAddress(buffer!, CVPixelBufferLockFlags(rawValue: 0))
return buffer
}
func createCMSampleBuffer(presentationTimeStamp: CMTime, duration:CMTime, decodeTimeStamp:CMTime ) -> CMSampleBuffer? {
let pixelBuffer = cvPixelBuffer
var newSampleBuffer: CMSampleBuffer?
var info = CMSampleTimingInfo()
var videoInfo: CMVideoFormatDescription?
info.presentationTimeStamp = presentationTimeStamp
info.duration = duration
info.decodeTimeStamp = CMTime.invalid
CMVideoFormatDescriptionCreateForImageBuffer(allocator: nil, imageBuffer: pixelBuffer!, formatDescriptionOut: &videoInfo)
CMSampleBufferCreateForImageBuffer(allocator: kCFAllocatorDefault,
imageBuffer: pixelBuffer!,
dataReady: true,
makeDataReadyCallback: nil,
refcon: nil,
formatDescription: videoInfo!,
sampleTiming: &info,
sampleBufferOut: &newSampleBuffer)
return newSampleBuffer!
}
}
Note: I am using Hashinkit but with my custom camera preview layer not attaching it to MTHKView becuase with view and camera both usage its getting heat warnings on Hashinkit
Related
Right now i am developing one module in that module i need to create video from array CGImage and while doing that processing my application get crashed at some point , i am not able to figure out exact reason behind that crash.
can anyone please suggest me i am going in right direction or not , should i convert [CGImage] to video or do i need to choose another approach.
i also tried to convert CGImage to UIImage and tried to create video but still facing same issue.
i am getting image data in [UInt8] data so what would be the correct approach converting the image formate and create video ?
In order to create video from [CGImage] following below approach.
I am converting [UInt8] data to CGImage using CGDataProvider and convert CGImage to UIImage. I have array of image and collect UIImage and then merge images and create video.
Here my code to convert CGImage from data.
private(set) var data: [UInt8]
var cgImage: CGImage? {
let colorSpaceRef = CGColorSpaceCreateDeviceRGB()
let bitsPerComponent = 8
let bitsPerPixel = channels * bitsPerComponent
let bytesPerRow = channels * width
let totalBytes = height * bytesPerRow
let bitmapInfo = CGBitmapInfo(rawValue: channels == 3 ? CGImageAlphaInfo.none.rawValue : CGImageAlphaInfo.last.rawValue)
let provider = CGDataProvider( dataInfo: nil,
data: data,
size: totalBytes,
releaseData: {_, _, _ in })!
return CGImage(width: width,
height: height,
bitsPerComponent: bitsPerComponent,
bitsPerPixel: bitsPerPixel,
bytesPerRow: bytesPerRow,
space: colorSpaceRef,
bitmapInfo: bitmapInfo,
provider: provider,
decode: nil,
shouldInterpolate: false,
intent: CGColorRenderingIntent.perceptual)
}
My app is getting crash here in this function, when i start frequent image drawing to context
(context!.draw(cgImage, in: CGRect(x: 0, y: 0, width: frameWidth,
height: frameHeight)))
If i use number of images from bundle and create video using this code its working fine. When i use created CGImage from [UInt8] data, it started getting crash after writing 3-4 images.
func newPixelBufferFrom(cgImage:CGImage) -> CVPixelBuffer?{
autoreleasepool {
let options:[String: Any] = [kCVPixelBufferCGImageCompatibilityKey as String: true, kCVPixelBufferCGBitmapContextCompatibilityKey as String: true]
var pxbuffer:CVPixelBuffer?
let frameWidth = self.videoSettings[AVVideoWidthKey] as! Int
let frameHeight = self.videoSettings[AVVideoHeightKey] as! Int
let status = CVPixelBufferCreate(kCFAllocatorDefault, frameWidth, frameHeight, kCVPixelFormatType_32ARGB, options as CFDictionary?, &pxbuffer)
assert(status == kCVReturnSuccess && pxbuffer != nil, "newPixelBuffer failed")
CVPixelBufferLockBaseAddress(pxbuffer!, CVPixelBufferLockFlags(rawValue: 0))
let pxdata = CVPixelBufferGetBaseAddress(pxbuffer!)
let rgbColorSpace = CGColorSpaceCreateDeviceRGB()
let context = CGContext(data: pxdata, width: frameWidth, height: frameHeight, bitsPerComponent: 8, bytesPerRow: CVPixelBufferGetBytesPerRow(pxbuffer!), space: rgbColorSpace, bitmapInfo: CGImageAlphaInfo.noneSkipFirst.rawValue)
assert(context != nil, "context is nil")
context!.concatenate(CGAffineTransform.identity)
context!.draw(cgImage
, in: CGRect(x: 0, y: 0, width: frameWidth, height: frameHeight))
CVPixelBufferUnlockBaseAddress(pxbuffer!, CVPixelBufferLockFlags(rawValue: 0))
return pxbuffer
}
here, i am using below code to create video from array of images.
typealias CXEMovieMakerCompletion = (URL) -> Void
typealias CXEMovieMakerUIImageExtractor = (AnyObject) -> UIImage?
public class CXEImagesToVideo: NSObject{
var assetWriter:AVAssetWriter!
var writeInput:AVAssetWriterInput!
var bufferAdapter:AVAssetWriterInputPixelBufferAdaptor!
var videoSettings:[String : Any]!
var frameTime:CMTime!
var fileURL:URL!
var completionBlock: CXEMovieMakerCompletion?
var movieMakerUIImageExtractor:CXEMovieMakerUIImageExtractor?
public class func videoSettings(codec:String, width:Int, height:Int) -> [String: Any]{
if(Int(width) % 16 != 0){
print("warning: video settings width must be divisible by 16")
}
let videoSettings:[String: Any] = [AVVideoCodecKey: AVVideoCodecType.h264,
AVVideoWidthKey: width,
AVVideoHeightKey: height]
return videoSettings
}
public init(videoSettings: [String: Any],frameTime: CMTime) {
super.init()
self.frameTime = frameTime
let paths = NSSearchPathForDirectoriesInDomains(.documentDirectory, .userDomainMask, true)
let tempPath = paths[0] + "/exprotvideo1.mp4"
if(FileManager.default.fileExists(atPath: tempPath)){
guard (try? FileManager.default.removeItem(atPath: tempPath)) != nil else {
print("remove path failed")
return
}
}
self.fileURL = URL(fileURLWithPath: tempPath)
self.assetWriter = try! AVAssetWriter(url: self.fileURL, fileType: AVFileType.mp4)
self.videoSettings = videoSettings
self.writeInput = AVAssetWriterInput(mediaType: AVMediaType.video, outputSettings: videoSettings)
assert(self.assetWriter.canAdd(self.writeInput), "add failed")
self.assetWriter.add(self.writeInput)
let bufferAttributes:[String: Any] = [kCVPixelBufferPixelFormatTypeKey as String: Int(kCVPixelFormatType_32ARGB)]
self.bufferAdapter = AVAssetWriterInputPixelBufferAdaptor(assetWriterInput: self.writeInput, sourcePixelBufferAttributes: bufferAttributes)
self.frameTime = CMTimeMake(value: 1, timescale: 10)
}
func createMovieFrom(urls: [URL], withCompletion: #escaping CXEMovieMakerCompletion){
self.createMovieFromSource(images: urls as [AnyObject], extractor:{(inputObject:AnyObject) ->UIImage? in
return UIImage(data: try! Data(contentsOf: inputObject as! URL))}, withCompletion: withCompletion)
}
func createMovieFrom(images: [UIImage], withCompletion: #escaping CXEMovieMakerCompletion){
DispatchQueue.main.async {
self.createMovieFromSource(images: images, extractor: {(inputObject:AnyObject) -> UIImage? in
return inputObject as? UIImage}, withCompletion: withCompletion)
}
}
func imageFromLayer(layer:CALayer) -> UIImage {
UIGraphicsBeginImageContextWithOptions(layer.frame.size, layer.isOpaque, 0)
layer.render(in: UIGraphicsGetCurrentContext()!)
let outputImage = UIGraphicsGetImageFromCurrentImageContext()
UIGraphicsEndImageContext()
return outputImage!
}
func createMovieFromSource(images: [AnyObject], extractor: #escaping CXEMovieMakerUIImageExtractor, withCompletion: #escaping CXEMovieMakerCompletion){
self.completionBlock = withCompletion
self.assetWriter.startWriting()
self.assetWriter.startSession(atSourceTime: CMTime.zero)
let mediaInputQueue = DispatchQueue.init(label: "Main") // DispatchQueue(label: "mediaInputQueue")
var i = 0
let frameNumber = images.count
self.writeInput.requestMediaDataWhenReady(on: mediaInputQueue){
while(true){
if(i >= frameNumber){
break
}
if (self.writeInput.isReadyForMoreMediaData){
var sampleBuffer:CVPixelBuffer?
autoreleasepool{
let temp = images[i]
let img = extractor(temp)
if img == nil{
i += 1
print("Warning: counld not extract one of the frames")
//continue
}
sampleBuffer = self.newPixelBufferFrom(cgImage: temp.cgImage!)
}
if (sampleBuffer != nil){
if(i == 0){
self.bufferAdapter.append(sampleBuffer!, withPresentationTime: CMTime.zero)
}else{
let value = i - 1
let lastTime = CMTimeMake(value: Int64(value), timescale: self.frameTime.timescale)
let presentTime = CMTimeAdd(lastTime, self.frameTime)
self.bufferAdapter.append(sampleBuffer!, withPresentationTime: presentTime)
}
i = i + 1
}
}
}
self.writeInput.markAsFinished()
self.assetWriter.finishWriting {
DispatchQueue.main.sync {
self.completionBlock!(self.fileURL)
}
}
}
}
func newPixelBufferFrom(cgImage:CGImage) -> CVPixelBuffer?{
autoreleasepool {
let options:[String: Any] = [kCVPixelBufferCGImageCompatibilityKey as String: true, kCVPixelBufferCGBitmapContextCompatibilityKey as String: true]
var pxbuffer:CVPixelBuffer?
let frameWidth = self.videoSettings[AVVideoWidthKey] as! Int
let frameHeight = self.videoSettings[AVVideoHeightKey] as! Int
let status = CVPixelBufferCreate(kCFAllocatorDefault, frameWidth, frameHeight, kCVPixelFormatType_32ARGB, options as CFDictionary?, &pxbuffer)
assert(status == kCVReturnSuccess && pxbuffer != nil, "newPixelBuffer failed")
CVPixelBufferLockBaseAddress(pxbuffer!, CVPixelBufferLockFlags(rawValue: 0))
let pxdata = CVPixelBufferGetBaseAddress(pxbuffer!)
let rgbColorSpace = CGColorSpaceCreateDeviceRGB()
let context = CGContext(data: pxdata, width: frameWidth, height: frameHeight, bitsPerComponent: 8, bytesPerRow: CVPixelBufferGetBytesPerRow(pxbuffer!), space: rgbColorSpace, bitmapInfo: CGImageAlphaInfo.noneSkipFirst.rawValue)
// CGImageAlphaInfo.noneSkipFirst.rawValue
assert(context != nil, "context is nil")
// context?.clear(CGRect(x: 0, y: 0, width: frameWidth, height: frameHeight))
context!.concatenate(CGAffineTransform.identity)
context!.draw(cgImage
, in: CGRect(x: 0, y: 0, width: frameWidth, height: frameHeight))
CVPixelBufferUnlockBaseAddress(pxbuffer!, CVPixelBufferLockFlags(rawValue: 0))
return pxbuffer
}
}
}
I have a tflite model and I want to run model using ARKit session captured image. It's showing source pixel format is invalid. I was able to run tflite model using AVCapture session:
// Not working with ARKit frame
func session(_ session: ARSession, didUpdate frame: ARFrame) {
guard currentBuffer == nil, case .normal = frame.camera.trackingState else { return }
runModel(with : currentBuffer)
}
// Working fine with AVCapture session
func captureOutput(_ output: AVCaptureOutput, didOutput sampleBuffer: CMSampleBuffer, from connection: AVCaptureConnection) {
let pixelBuffer = CMSampleBufferGetImageBuffer(sampleBuffer)
runModel(with : pixelBuffer)
}
let context = CIContext()
func session(_ session: ARSession, didUpdate frame: ARFrame) {
guard currentBuffer == nil, case .normal = frame.camera.trackingState else {
return
}
guard let scaledPixelBuffer = CIImage(cvPixelBuffer: frame.capturedImage)
.oriented(.right)
.resize(size: CGSize(width: 256, height: 256))
.toPixelBuffer(context: context)
else {
return
}
runModel(with : scaledPixelBuffer)
}
extension CIImage {
func resize(size : CGSize) -> CIImage {
let scale = min(size.width, size.height) / min(self.extent.size.width, self.extent.size.height)
let resizedImage = self.transformed(by: CGAffineTransform(scaleX: scale, y: scale))
let width = resizedImage.extent.size.width
let height = resizedImage.extent.size.height
let xOffset = (CGFloat(width) - size.width) / 2.0
let yOffset = (CGFloat(height) - size.height) / 2.0
let rect = CGRect(x: xOffset, y: yOffset, width: size.width, height: size.height)
return resizedImage
.clamped(to: rect)
.cropped(to: CGRect(x: 0, y: 0, width: size.width, height: size.height))
}
func toPixelBuffer(context : CIContext, size inSize:CGSize? = nil, gray : Bool = true) -> CVPixelBuffer? {
let attributes = [
kCVPixelBufferCGImageCompatibilityKey: kCFBooleanTrue,
kCVPixelBufferCGBitmapContextCompatibilityKey: kCFBooleanTrue ] as CFDictionary
var nullablePixelBuffer : CVPixelBuffer? = nil
let status = CVPixelBufferCreate(kCFAllocatorDefault, Int(self.extent.size.width), Int(self.extent.size.height), kCVPixelFormatType_32BGRA, attributes, &nullablePixelBuffer)
guard status == kCVReturnSuccess, let pixelBuffer = nullablePixelBuffer else {
return nil
}
CVPixelBufferLockBaseAddress(pixelBuffer, CVPixelBufferLockFlags(rawValue: 0))
context.render(self, to: pixelBuffer, bounds: CGRect(x: 0, y: 0, width: self.extent.size.width, height: self.extent.size.height), colorSpace: gray ? CGColorSpaceCreateDeviceGray() : self.colorSpace)
CVPixelBufferUnlockBaseAddress(pixelBuffer, CVPixelBufferLockFlags(rawValue: 0))
return pixelBuffer
}
}
Some questions address how to convert a CMSampleBuffer to a UIImage, but there are no answers on how to do the reverse, i.e., convert UIImage to CMSampleBuffer.
This question is different from similar ones because the code below provides a starting point for converting a UIImage to a CVPixelBuffer, which hopefully someone with more AVFoundation expertise can help fix to convert to a CMSampleBuffer.
func convertImageToBuffer(from image: UIImage) -> CVPixelBuffer? {
let attrs = [
String(kCVPixelBufferCGImageCompatibilityKey) : kCFBooleanTrue,
String(kCVPixelBufferCGBitmapContextCompatibilityKey) : kCFBooleanTrue
] as [String : Any]
var buffer : CVPixelBuffer?
let status = CVPixelBufferCreate(kCFAllocatorDefault, Int(image.size.width), Int(image.size.height), kCVPixelFormatType_32ARGB, attrs as CFDictionary, &buffer)
guard (status == kCVReturnSuccess) else {
return nil
}
CVPixelBufferLockBaseAddress(buffer!, CVPixelBufferLockFlags(rawValue: 0))
let pixelData = CVPixelBufferGetBaseAddress(buffer!)
let rgbColorSpace = CGColorSpaceCreateDeviceRGB()
let context = CGContext(data: pixelData, width: Int(image.size.width), height: Int(image.size.height), bitsPerComponent: 8, bytesPerRow: CVPixelBufferGetBytesPerRow(buffer!), space: rgbColorSpace, bitmapInfo: CGImageAlphaInfo.noneSkipFirst.rawValue)
context?.translateBy(x: 0, y: image.size.height)
context?.scaleBy(x: 1.0, y: -1.0)
UIGraphicsPushContext(context!)
image.draw(in: CGRect(x: 0, y: 0, width: image.size.width, height: image.size.height))
UIGraphicsPopContext()
CVPixelBufferUnlockBaseAddress(buffer!, CVPixelBufferLockFlags(rawValue: 0))
return buffer
}
You are halfway done. Just convert CVPixelBuffer to CMSampleBuffer:
extension UIImage {
var cvPixelBuffer: CVPixelBuffer? {
let attrs = [
String(kCVPixelBufferCGImageCompatibilityKey): kCFBooleanTrue,
String(kCVPixelBufferCGBitmapContextCompatibilityKey): kCFBooleanTrue
] as [String: Any]
var buffer: CVPixelBuffer?
let status = CVPixelBufferCreate(kCFAllocatorDefault, Int(self.size.width), Int(self.size.height), kCVPixelFormatType_32ARGB, attrs as CFDictionary, &buffer)
guard status == kCVReturnSuccess else {
return nil
}
CVPixelBufferLockBaseAddress(buffer!, CVPixelBufferLockFlags(rawValue: 0))
let pixelData = CVPixelBufferGetBaseAddress(buffer!)
let rgbColorSpace = CGColorSpaceCreateDeviceRGB()
let context = CGContext(data: pixelData, width: Int(self.size.width), height: Int(self.size.height), bitsPerComponent: 8, bytesPerRow: CVPixelBufferGetBytesPerRow(buffer!), space: rgbColorSpace, bitmapInfo: CGImageAlphaInfo.noneSkipFirst.rawValue)
context?.translateBy(x: 0, y: self.size.height)
context?.scaleBy(x: 1.0, y: -1.0)
UIGraphicsPushContext(context!)
self.draw(in: CGRect(x: 0, y: 0, width: self.size.width, height: self.size.height))
UIGraphicsPopContext()
CVPixelBufferUnlockBaseAddress(buffer!, CVPixelBufferLockFlags(rawValue: 0))
return buffer
}
func createCMSampleBuffer() -> CMSampleBuffer? {
let pixelBuffer = cvPixelBuffer
var newSampleBuffer: CMSampleBuffer?
var timimgInfo: CMSampleTimingInfo?
var videoInfo: CMVideoFormatDescription?
CMVideoFormatDescriptionCreateForImageBuffer(allocator: nil, imageBuffer: pixelBuffer!, formatDescriptionOut: &videoInfo)
CMSampleBufferCreateForImageBuffer(allocator: kCFAllocatorDefault,
imageBuffer: pixelBuffer!,
dataReady: true,
makeDataReadyCallback: nil,
refcon: nil,
formatDescription: videoInfo!,
sampleTiming: &timimgInfo,
sampleBufferOut: &newSampleBuffer)
return newSampleBuffer!
}
}
I'm handling ReplayKit2 in iOS, for some reasons I need to rotate CMSampleBuffer from portrait to landscape, I found the result is not correct.
What I miss ?
this is original sample buffer
this is actual output buffer
width & height are dimensions of sampleBuffer
func rotation(sampleBuffer: CMSampleBuffer, width: Int, height: Int) -> CMSampleBuffer {
//create pixelbuffer from the delegate method samplebuffer
let pixelBuffer = CMSampleBufferGetImageBuffer(sampleBuffer)!
CVPixelBufferLockBaseAddress(pixelBuffer, CVPixelBufferLockFlags(rawValue: 0))
//create CI image from the buffer
let image = CIImage(cvImageBuffer: pixelBuffer)
let extent = CGRect(x: 0, y: 0, width: width, height: height)
var tx = CGAffineTransform(translationX: extent.midX, y: extent.midY)
tx = tx.rotated(by: CGFloat(Double.pi / 2))
tx = tx.translatedBy(x: -extent.midX, y: -extent.midY)
var transformImage = CIFilter(
name: "CIAffineTransform",
withInputParameters: [
kCIInputImageKey: image,
kCIInputTransformKey: NSValue.init(cgAffineTransform: tx)])!.outputImage!
//create empty pixelbuffer
var newPixelBuffer : CVPixelBuffer? = nil
CVPixelBufferCreate(kCFAllocatorDefault,
width,
height,
kCVPixelFormatType_32BGRA,
nil,
&newPixelBuffer)
//render the context to the new pixelbuffer, context is a global
//CIContext variable. creating a new one each frame is too CPU intensive
self.ciContext.render(transformImage, to: newPixelBuffer!)
//finally, write this to the pixelbufferadaptor
CVPixelBufferUnlockBaseAddress(pixelBuffer,CVPixelBufferLockFlags(rawValue: 0))
var videoInfo: CMVideoFormatDescription?
CMVideoFormatDescriptionCreateForImageBuffer(kCFAllocatorDefault, newPixelBuffer!, &videoInfo)
var sampleTimingInfo = CMSampleTimingInfo(duration: CMSampleBufferGetDuration(sampleBuffer), presentationTimeStamp: CMSampleBufferGetPresentationTimeStamp(sampleBuffer), decodeTimeStamp: CMSampleBufferGetDecodeTimeStamp(sampleBuffer))
var newSampleBuffer: CMSampleBuffer?
CMSampleBufferCreateForImageBuffer(kCFAllocatorDefault, newPixelBuffer!, true, nil, nil, videoInfo!, &sampleTimingInfo, &newSampleBuffer)
return newSampleBuffer!
}
just found a very sweet method in iOS 11!
/* Returns a new image representing the original image transformeded for the given CGImagePropertyOrientation */
#available(iOS 11.0, *)
open func oriented(_ orientation: CGImagePropertyOrientation) -> CIImage
May be it will be useful
func rotate(_ sampleBuffer: CMSampleBuffer) -> CVPixelBuffer? {
guard let pixelBuffer = CMSampleBufferGetImageBuffer(sampleBuffer) else {
return nil
}
var newPixelBuffer: CVPixelBuffer?
let error = CVPixelBufferCreate(kCFAllocatorDefault,
CVPixelBufferGetHeight(pixelBuffer),
CVPixelBufferGetWidth(pixelBuffer),
kCVPixelFormatType_420YpCbCr8BiPlanarFullRange,
nil,
&newPixelBuffer)
guard error == kCVReturnSuccess else {
return nil
}
let ciImage = CIImage(cvPixelBuffer: pixelBuffer).oriented(.right)
let context = CIContext(options: nil)
context.render(ciImage, to: newPixelBuffer!)
return newPixelBuffer
}
I am trying to get Apple's sample Core ML Models that were demoed at the 2017 WWDC to function correctly. I am using the GoogLeNet to try and classify images (see the Apple Machine Learning Page). The model takes a CVPixelBuffer as an input. I have an image called imageSample.jpg that I'm using for this demo. My code is below:
var sample = UIImage(named: "imageSample")?.cgImage
let bufferThree = getCVPixelBuffer(sample!)
let model = GoogLeNetPlaces()
guard let output = try? model.prediction(input: GoogLeNetPlacesInput.init(sceneImage: bufferThree!)) else {
fatalError("Unexpected runtime error.")
}
print(output.sceneLabel)
I am always getting the unexpected runtime error in the output rather than an image classification. My code to convert the image is below:
func getCVPixelBuffer(_ image: CGImage) -> CVPixelBuffer? {
let imageWidth = Int(image.width)
let imageHeight = Int(image.height)
let attributes : [NSObject:AnyObject] = [
kCVPixelBufferCGImageCompatibilityKey : true as AnyObject,
kCVPixelBufferCGBitmapContextCompatibilityKey : true as AnyObject
]
var pxbuffer: CVPixelBuffer? = nil
CVPixelBufferCreate(kCFAllocatorDefault,
imageWidth,
imageHeight,
kCVPixelFormatType_32ARGB,
attributes as CFDictionary?,
&pxbuffer)
if let _pxbuffer = pxbuffer {
let flags = CVPixelBufferLockFlags(rawValue: 0)
CVPixelBufferLockBaseAddress(_pxbuffer, flags)
let pxdata = CVPixelBufferGetBaseAddress(_pxbuffer)
let rgbColorSpace = CGColorSpaceCreateDeviceRGB();
let context = CGContext(data: pxdata,
width: imageWidth,
height: imageHeight,
bitsPerComponent: 8,
bytesPerRow: CVPixelBufferGetBytesPerRow(_pxbuffer),
space: rgbColorSpace,
bitmapInfo: CGImageAlphaInfo.premultipliedFirst.rawValue)
if let _context = context {
_context.draw(image, in: CGRect.init(x: 0, y: 0, width: imageWidth, height: imageHeight))
}
else {
CVPixelBufferUnlockBaseAddress(_pxbuffer, flags);
return nil
}
CVPixelBufferUnlockBaseAddress(_pxbuffer, flags);
return _pxbuffer;
}
return nil
}
I got this code from a previous StackOverflow post (last answer here). I recognize that the code may not be correct, but I have no idea of how to do this myself. I believe that this is the section that contains the error. The model calls for the following type of input: Image<RGB,224,224>
You don't need to do a bunch of image mangling yourself to use a Core ML model with an image — the new Vision framework can do that for you.
import Vision
import CoreML
let model = try VNCoreMLModel(for: MyCoreMLGeneratedModelClass().model)
let request = VNCoreMLRequest(model: model, completionHandler: myResultsMethod)
let handler = VNImageRequestHandler(url: myImageURL)
handler.perform([request])
func myResultsMethod(request: VNRequest, error: Error?) {
guard let results = request.results as? [VNClassificationObservation]
else { fatalError("huh") }
for classification in results {
print(classification.identifier, // the scene label
classification.confidence)
}
}
The WWDC17 session on Vision should have a bit more info — it's tomorrow afternoon.
You can use a pure CoreML, but you should resize an image to (224,224)
DispatchQueue.global(qos: .userInitiated).async {
// Resnet50 expects an image 224 x 224, so we should resize and crop the source image
let inputImageSize: CGFloat = 224.0
let minLen = min(image.size.width, image.size.height)
let resizedImage = image.resize(to: CGSize(width: inputImageSize * image.size.width / minLen, height: inputImageSize * image.size.height / minLen))
let cropedToSquareImage = resizedImage.cropToSquare()
guard let pixelBuffer = cropedToSquareImage?.pixelBuffer() else {
fatalError()
}
guard let classifierOutput = try? self.classifier.prediction(image: pixelBuffer) else {
fatalError()
}
DispatchQueue.main.async {
self.title = classifierOutput.classLabel
}
}
// ...
extension UIImage {
func resize(to newSize: CGSize) -> UIImage {
UIGraphicsBeginImageContextWithOptions(CGSize(width: newSize.width, height: newSize.height), true, 1.0)
self.draw(in: CGRect(x: 0, y: 0, width: newSize.width, height: newSize.height))
let resizedImage = UIGraphicsGetImageFromCurrentImageContext()!
UIGraphicsEndImageContext()
return resizedImage
}
func cropToSquare() -> UIImage? {
guard let cgImage = self.cgImage else {
return nil
}
var imageHeight = self.size.height
var imageWidth = self.size.width
if imageHeight > imageWidth {
imageHeight = imageWidth
}
else {
imageWidth = imageHeight
}
let size = CGSize(width: imageWidth, height: imageHeight)
let x = ((CGFloat(cgImage.width) - size.width) / 2).rounded()
let y = ((CGFloat(cgImage.height) - size.height) / 2).rounded()
let cropRect = CGRect(x: x, y: y, width: size.height, height: size.width)
if let croppedCgImage = cgImage.cropping(to: cropRect) {
return UIImage(cgImage: croppedCgImage, scale: 0, orientation: self.imageOrientation)
}
return nil
}
func pixelBuffer() -> CVPixelBuffer? {
let width = self.size.width
let height = self.size.height
let attrs = [kCVPixelBufferCGImageCompatibilityKey: kCFBooleanTrue,
kCVPixelBufferCGBitmapContextCompatibilityKey: kCFBooleanTrue] as CFDictionary
var pixelBuffer: CVPixelBuffer?
let status = CVPixelBufferCreate(kCFAllocatorDefault,
Int(width),
Int(height),
kCVPixelFormatType_32ARGB,
attrs,
&pixelBuffer)
guard let resultPixelBuffer = pixelBuffer, status == kCVReturnSuccess else {
return nil
}
CVPixelBufferLockBaseAddress(resultPixelBuffer, CVPixelBufferLockFlags(rawValue: 0))
let pixelData = CVPixelBufferGetBaseAddress(resultPixelBuffer)
let rgbColorSpace = CGColorSpaceCreateDeviceRGB()
guard let context = CGContext(data: pixelData,
width: Int(width),
height: Int(height),
bitsPerComponent: 8,
bytesPerRow: CVPixelBufferGetBytesPerRow(resultPixelBuffer),
space: rgbColorSpace,
bitmapInfo: CGImageAlphaInfo.noneSkipFirst.rawValue) else {
return nil
}
context.translateBy(x: 0, y: height)
context.scaleBy(x: 1.0, y: -1.0)
UIGraphicsPushContext(context)
self.draw(in: CGRect(x: 0, y: 0, width: width, height: height))
UIGraphicsPopContext()
CVPixelBufferUnlockBaseAddress(resultPixelBuffer, CVPixelBufferLockFlags(rawValue: 0))
return resultPixelBuffer
}
}
The expected image size for inputs you can find in the mimodel file:
A demo project that uses both pure CoreML and Vision variants you can find here: https://github.com/handsomecode/iOS11-Demos/tree/coreml_vision/CoreML/CoreMLDemo
If the input is UIImage, rather than an URL, and you want to use VNImageRequestHandler, you can use CIImage.
func updateClassifications(for image: UIImage) {
let orientation = CGImagePropertyOrientation(image.imageOrientation)
guard let ciImage = CIImage(image: image) else { return }
let handler = VNImageRequestHandler(ciImage: ciImage, orientation: orientation)
}
From Classifying Images with Vision and Core ML