vImageConvolve_Planar16F(_:_:_:_:_:_:_:_:_:_:)

Convolves a floating-point 16-bit planar image by a 2D kernel.

Declaration

func vImageConvolve_Planar16F(_ src: UnsafePointer<vImage_Buffer>, _ dest: UnsafePointer<vImage_Buffer>, _ tempBuffer: UnsafeMutableRawPointer!, _ srcOffsetToROI_X: vImagePixelCount, _ srcOffsetToROI_Y: vImagePixelCount, _ kernel: UnsafePointer<Float>!, _ kernel_height: UInt32, _ kernel_width: UInt32, _ backgroundColor: Pixel_16F, _ flags: vImage_Flags) -> vImage_Error

Parameters

src:
A pointer to a vImage buffer structure that contains data for the source image.
dest:
A pointer to a vImage buffer data structure. You’re responsible for filling out the height, width, and rowBytes fields of this structure, and for allocating a data buffer of the appropriate size. On return, the data buffer this structure points to contains the destination image data. When you no longer need the data buffer, you need to deallocate the memory. The size (number of rows and number of columns) of the destination buffer also specifies the size of the region of interest in the source buffer.
tempBuffer:
A pointer to workspace memory the function uses as it operates on an image. Pass nil to instruct the function to allocate, use, and then free its own temporary buffer.
srcOffsetToROI_X:
The horizontal offset, in pixels, to the upper-left pixel of the region of interest within the source image.
srcOffsetToROI_Y:
The vertical offset, in pixels, to the upper-left pixel of the region of interest within the source image.
kernel:
A pointer to the convolution kernel data, which needs to be a packed array without any padding.
kernel_height:
The height of the kernel in pixels. This value needs to be odd.
kernel_width:
The width of the kernel in pixels. This value needs to be odd.
backgroundColor:
A background color. If you supply a color, you need to also set the kvImageBackgroundColorFill flag; otherwise, the function ignores the color.
flags:
The options to use when performing the operation. If your code implements its own tiling or its own multithreading, pass Kvimagedonottile.
To instruct the function to return the minimum size of the workspace memory, set the Kvimagegettempbuffersize flag.
Pass one of the following flags to specify how vImage handles pixel locations beyond the edge of the source image: Kvimagecopyinplace, Kvimagetruncatekernel, Kvimagebackgroundcolorfill, or Kvimageedgeextend.
If you want vImage to use faster but lower-precision internal arithmetic, set the Kvimageusefp16accumulator flag.

Return Value

kvImageNoError; otherwise, a negative value indicates one of the error codes that Data Types and Constants describes, and a positive value indicates the required size for the temporary buffer.

Discussion

Optimize performance with temporary buffers

This function uses a multiple-pass algorithm that saves intermediate pixel values between passes. In some cases, the destination buffer may not be large enough to store that intermediate data, so the operation requires additional storage.

Pass nil to the tempBuffer parameter to have vImage create and manage this temporary storage for you.

In cases where your code calls the function frequently (for example, when processing video), create and manage this temporary buffer yourself and reuse it across function calls. Reusing a buffer avoids vImage allocating the temporary storage with each call.

To use your own temporary buffer, first call the function with the same values for all other parameters that you intend to use for subsequent calls. In addition, pass the kvImageGetTempBufferSize flag. The kvImageGetTempBufferSize instructs the function not to perform any processing, and to return a positive value that represents the minimum size, in bytes, of the temporary buffer. A negative return value represents an error.

After you allocate the memory for the temporary buffer, pass that to the tempBuffer parameter for subsequent calls to the function, and don’t pass the kvImageGetTempBufferSize flag.

You can use the same workspace memory for a group of images that are different sizes. To do this, create a vImage buffer structure with a size that shares the maximum width and maximum height of the images that you’re working with. For example, to create a workspace memory that’s suitable for 4x4, 5x4, and 4x5 images, pass a buffer with the size 5x5.