Swift 使用加速器规范化视差图像
我有一个视差图像,我正在使用下面的示例代码对其进行规范化,但速度非常慢。 我需要使用一些加速器,比如定制的CIFilter或其他技术,但我不知道怎么做?我目前正在用CIContext()运行代码,它正在CPU上运行(不确定)。有没有一种方法可以在GPU上运行它并在没有定制CIfilter的情况下加速? 以下是当前代码:Swift 使用加速器规范化视差图像,swift,core-image,disparity-mapping,Swift,Core Image,Disparity Mapping,我有一个视差图像,我正在使用下面的示例代码对其进行规范化,但速度非常慢。 我需要使用一些加速器,比如定制的CIFilter或其他技术,但我不知道怎么做?我目前正在用CIContext()运行代码,它正在CPU上运行(不确定)。有没有一种方法可以在GPU上运行它并在没有定制CIfilter的情况下加速? 以下是当前代码: extension CVPixelBuffer { func normalize() { let width = CVPixelBufferGetWidth(se
extension CVPixelBuffer {
func normalize() {
let width = CVPixelBufferGetWidth(self)
let height = CVPixelBufferGetHeight(self)
CVPixelBufferLockBaseAddress(self, CVPixelBufferLockFlags(rawValue: 0))
let baseAddr = CVPixelBufferGetBaseAddress(self)!
let floatBuffer = unsafeBitCast(CVPixelBufferGetBaseAddress(self), to: UnsafeMutablePointer<Float>.self)
var minPixel: Float = 1.0
var maxPixel: Float = 0.0
for y in 0 ..< height {
for x in 0 ..< width {
let pixel = floatBuffer[y * width + x]
minPixel = min(pixel, minPixel)
maxPixel = max(pixel, maxPixel)
}
}
let range = maxPixel - minPixel
for y in 0 ..< height {
for x in 0 ..< width {
let pixel = floatBuffer[y * width + x]
floatBuffer[y * width + x] = (pixel - minPixel) / range
}
}
CVPixelBufferUnlockBaseAddress(self, CVPixelBufferLockFlags(rawValue: 0))
}
}
扩展CVPixelBuffer{
func正规化(){
let width=CVPixelBufferGetWidth(自)
let height=CVPixelBufferGetHeight(自)
CVPixelBufferLockBaseAddress(自身,CVPixelBufferLockFlags(原始值:0))
让baseAddr=CVPixelBufferGetBaseAddress(self)!
让floatBuffer=unsafeBitCast(CVPixelBufferGetBaseAddress(self),to:UnsafeMutablePointer.self)
var minPixel:Float=1.0
var maxPixel:Float=0.0
对于y,0..<高度{
对于0中的x..<宽度{
设像素=浮动缓冲区[y*宽度+x]
最小像素=最小(像素,最小像素)
maxPixel=max(像素,maxPixel)
}
}
让范围=最大像素-最小像素
对于y,0..<高度{
对于0中的x..<宽度{
设像素=浮动缓冲区[y*宽度+x]
浮动缓冲区[y*宽度+x]=(像素-最小像素)/范围
}
}
CVPixelBufferUnlockBaseAddress(self,CVPixelBufferLockFlags(rawValue:0))
}
}
对于您的用例,vImage可能是最好的选择。见附件中的选项3
在核心映像中也有这样做的方法。我可以想象使用
CIAreaMinMax
过滤器获取极值,然后使用一些巧妙的混合进行规范化。如果您愿意,我可以详细说明。您的像素值为浮点值,因此您也可以使用vDSP
计算极值,然后:
floatBuffer[y * width + x] = (pixel - minPixel) / range
可以替换为(需要乘以范围的倒数)
查看此计算中的哪一项可能也很有用:
m = sum(A[n], 0 <= n < N) / N;
d = sqrt(sum(A[n]**2, 0 <= n < N) / N - m**2);
if (C)
{
// Normalize.
for (n = 0; n < N; ++n)
C[n] = (A[n] - m) / d;
}
m=sum(A[n],0我使用加速框架vDSP向量函数来规范化视差。有关工作演示,请参阅gitHub中的修改
下面是两个函数中的相关代码
extension CVPixelBuffer {
func vectorNormalize( targetVector: UnsafeMutableBufferPointer<Float>) -> [Float] {
// range = max - min
// normalized to 0..1 is (pixel - minPixel) / range
// see Documentation "Using vDSP for Vector-based Arithmetic" in vDSP under system "Accelerate" documentation
// see also the Accelerate documentation section 'Vector extrema calculation'
// Maximium static func maximum<U>(U) -> Float
// Returns the maximum element of a single-precision vector.
//static func minimum<U>(U) -> Float
// Returns the minimum element of a single-precision vector.
let maxValue = vDSP.maximum(targetVector)
let minValue = vDSP.minimum(targetVector)
let range = maxValue - minValue
let negMinValue = -minValue
let subtractVector = vDSP.add(negMinValue, targetVector)
// adding negative value is subtracting
let result = vDSP.divide(subtractVector, range)
return result
}
func setUpNormalize() -> CVPixelBuffer {
// grayscale buffer float32 ie Float
// return normalized CVPixelBuffer
CVPixelBufferLockBaseAddress(self,
CVPixelBufferLockFlags(rawValue: 0))
let width = CVPixelBufferGetWidthOfPlane(self, 0)
let height = CVPixelBufferGetHeightOfPlane(self, 0)
let count = width * height
let bufferBaseAddress = CVPixelBufferGetBaseAddressOfPlane(self, 0)
// UnsafeMutableRawPointer
let pixelBufferBase = unsafeBitCast(bufferBaseAddress, to: UnsafeMutablePointer<Float>.self)
let depthCopy = UnsafeMutablePointer<Float>.allocate(capacity: count)
depthCopy.initialize(from: pixelBufferBase, count: count)
let depthCopyBuffer = UnsafeMutableBufferPointer<Float>(start: depthCopy, count: count)
let normalizedDisparity = vectorNormalize(targetVector: depthCopyBuffer)
pixelBufferBase.initialize(from: normalizedDisparity, count: count)
// copy back the normalized map into the CVPixelBuffer
depthCopy.deallocate()
// depthCopyBuffer.deallocate()
CVPixelBufferUnlockBaseAddress(self, CVPixelBufferLockFlags(rawValue: 0))
return self
}
}
扩展CVPixelBuffer{
func vectorNormalize(targetVector:UnsafemtableBufferPointer)->[浮点]{
//范围=最大值-最小值
//标准化为0..1是(像素-最小像素)/范围
//请参阅vDSP中系统“加速”文档下的“使用vDSP进行矢量运算”文档
//另请参见加速文档部分“向量极值计算”
//最大静态函数最大值(U)->浮点
//返回单个精度向量的最大元素。
//静态函数最小值(U)->浮点
//返回单个精度向量的最小元素。
设maxValue=vDSP.maximum(targetVector)
设minValue=vDSP.minimum(targetVector)
让范围=最大值-最小值
设negMinValue=-minValue
让subtractVector=vDSP.add(negMinValue,targetVector)
//加负值等于减
让结果=vDSP.除法(减法向量,范围)
返回结果
}
func setUpNormalize()->CVPixelBuffer{
//灰度缓冲浮点32 ie浮点
//返回标准化的CVPixelBuffer
CVPixelBufferLockBaseAddress(自,
CVPixelBufferLockFlags(原始值:0))
let width=CVPixelBufferGetWidthOfPlane(self,0)
let height=CVPixelBufferGetHeightof平面(自,0)
让计数=宽度*高度
让bufferBaseAddress=CVPixelBufferGetBaseAddressOfPlane(self,0)
//不可女性化的指针
让pixelBufferBase=unsafeBitCast(bufferBaseAddress,to:UnsafeMutablePointer.self)
let depthCopy=UnsafeMutablePointer.allocate(容量:计数)
初始化(起始:pixelBufferBase,计数:计数)
让depthCopyBuffer=UnsafemeutableBufferPointer(开始:depthCopy,计数:计数)
让normalizedDisparity=vectorNormalize(targetVector:depthCopyBuffer)
pixelBufferBase.initialize(起始:normalizedDisparity,计数:计数)
//将规格化贴图复制回CVPixelBuffer
depthCopy.deallocate()
//depthCopyBuffer.deallocate()
CVPixelBufferUnlockBaseAddress(self,CVPixelBufferLockFlags(rawValue:0))
回归自我
}
}
谢谢您的回答。是的,如果您给出代码示例,这将非常有用。我试图让它与内置CI过滤器一起工作,但目前为止运气不佳。对于您的用例,使用vImage
或vDSP
很可能是更快的选择。谢谢。我会的。我认为这是一个正确的答案,尽管不再需要它了(我用CIContext在真实设备上运行了我的代码,速度足够快)。如果将来需要额外的速度,我会使用你的代码。再次感谢你!