# Convolution Apply a convolution kernel to an image. ## Discussion *Convolution* is a common image-processing technique that changes the value of a pixel according to the values of its surrounding pixels. Many common image filters, such as blurring, detecting edges, sharpening, and embossing, derive from convolution. *Kernels* form the basis of convolution operations. Kernels are arrays or matrices of weights that indicate the influence of a pixel’s neighbors on its final value. To calculate the value of each transformed pixel, a convolution operation adds the products of each surrounding pixel value with the corresponding kernel value. During a convolution operation, the kernel passes over every pixel in the image, repeating this procedure, and then applies the effect to the entire image. ## Topics ### Convolving an 8-bit image with 32-bit weights [`vImageConvolveFloatKernel_ARGB8888`](/documentation/Accelerate/vImageConvolveFloatKernel_ARGB8888(_:_:_:_:_:_:_:_:_:_:_:)) Convolves an 8-bit-per-channel, 4-channel interleaved image using 32-bit weights. ### Convolving with separable filter kernels [`vImageSepConvolve_Planar8`](/documentation/Accelerate/vImageSepConvolve_Planar8(_:_:_:_:_:_:_:_:_:_:_:_:)) Convolves an 8-bit planar image by separate horizontal and vertical separable kernels. [`vImageSepConvolve_Planar16U`](/documentation/Accelerate/vImageSepConvolve_Planar16U(_:_:_:_:_:_:_:_:_:_:_:_:)) Convolves an unsigned 16-bit planar image by separate horizontal and vertical separable kernels. [`vImageSepConvolve_Planar16F`](/documentation/Accelerate/vImageSepConvolve_Planar16F(_:_:_:_:_:_:_:_:_:_:_:_:)) Convolves a floating-point 16-bit planar image by separate horizontal and vertical separable kernels. [`vImageSepConvolve_PlanarF`](/documentation/Accelerate/vImageSepConvolve_PlanarF(_:_:_:_:_:_:_:_:_:_:_:_:)) Convolves a floating-point 32-bit planar image by separate horizontal and vertical separable kernels. [`vImageSepConvolve_Planar8to16U`](/documentation/Accelerate/vImageSepConvolve_Planar8to16U(_:_:_:_:_:_:_:_:_:_:_:_:_:)) Convolves an 8-bit planar image by separate horizontal and vertical separable kernels, and writes the result to an unsigned 16-bit planar destination. [`vImageSepConvolve_ARGB8888`](/documentation/Accelerate/vImageSepConvolve_ARGB8888(_:_:_:_:_:_:_:_:_:_:_:_:)) Convolves an 8-bit-per-channel, 4-channel interleaved image by separate horizontal and vertical separable kernels. ### Convolving without bias [`vImageConvolve_Planar8`](/documentation/Accelerate/vImageConvolve_Planar8(_:_:_:_:_:_:_:_:_:_:_:)) Convolves an 8-bit planar image by a 2D kernel and divides the pixel values by a divisor. [`vImageConvolve_Planar16F`](/documentation/Accelerate/vImageConvolve_Planar16F(_:_:_:_:_:_:_:_:_:_:)) Convolves a floating-point 16-bit planar image by a 2D kernel. [`vImageConvolve_PlanarF`](/documentation/Accelerate/vImageConvolve_PlanarF(_:_:_:_:_:_:_:_:_:_:)) Convolves a floating-point 32-bit planar image by a 2D kernel. [`vImageConvolve_ARGB8888`](/documentation/Accelerate/vImageConvolve_ARGB8888(_:_:_:_:_:_:_:_:_:_:_:)) Convolves an 8-bit-per-channel, 4-channel interleaved image by a 2D kernel and divides the pixel values by a divisor. [`vImageConvolve_ARGB16F`](/documentation/Accelerate/vImageConvolve_ARGB16F(_:_:_:_:_:_:_:_:_:_:)) Convolves a floating-point 16-bit-per-channel, 4-channel interleaved image by a 2D kernel, then divides the pixel values by a divisor. [`vImageConvolve_ARGBFFFF`](/documentation/Accelerate/vImageConvolve_ARGBFFFF(_:_:_:_:_:_:_:_:_:_:)) Convolves a floating-point 32-bit-per-channel, 4-channel interleaved image by a 2D kernel, then divides the pixel values by a divisor. ### Convolving with bias [`vImageConvolveWithBias_Planar8`](/documentation/Accelerate/vImageConvolveWithBias_Planar8(_:_:_:_:_:_:_:_:_:_:_:_:)) Convolves an 8-bit planar image by a 2D kernel and adds a bias. [`vImageConvolveWithBias_Planar16F`](/documentation/Accelerate/vImageConvolveWithBias_Planar16F(_:_:_:_:_:_:_:_:_:_:_:)) Convolves a floating-point 16-bit planar image by a 2D kernel and adds a bias. [`vImageConvolveWithBias_PlanarF`](/documentation/Accelerate/vImageConvolveWithBias_PlanarF(_:_:_:_:_:_:_:_:_:_:_:)) Convolves a floating-point 32-bit planar image by a 2D kernel and adds a bias. [`vImageConvolveWithBias_ARGB8888`](/documentation/Accelerate/vImageConvolveWithBias_ARGB8888(_:_:_:_:_:_:_:_:_:_:_:_:)) Convolves an 8-bit-per-channel, 4-channel interleaved image by a 2D kernel and adds a bias. [`vImageConvolveWithBias_ARGB16F`](/documentation/Accelerate/vImageConvolveWithBias_ARGB16F(_:_:_:_:_:_:_:_:_:_:_:)) Convolves a floating-point 16-bit-per-channel, 4-channel interleaved image by a 2D kernel and adds a bias. [`vImageConvolveWithBias_ARGBFFFF`](/documentation/Accelerate/vImageConvolveWithBias_ARGBFFFF(_:_:_:_:_:_:_:_:_:_:_:)) Convolves a floating-point 32-bit-per-channel, 4-channel interleaved image by a 2D kernel and adds a bias. ### Convolving with multiple kernels [`vImageConvolveMultiKernel_ARGB8888`](/documentation/Accelerate/vImageConvolveMultiKernel_ARGB8888(_:_:_:_:_:_:_:_:_:_:_:_:)) Convolves each channel of an 8-bit-per-channel, 4-channel interleaved image by one of the four 2D kernels. [`vImageConvolveMultiKernel_ARGBFFFF`](/documentation/Accelerate/vImageConvolveMultiKernel_ARGBFFFF(_:_:_:_:_:_:_:_:_:_:_:)) Convolves each channel of a floating-point 32-bit-per-channel, 4-channel interleaved image by one of the four 2D kernels. ### Convolving with high-speed box and tent filters [`vImageBoxConvolve_Planar8`](/documentation/Accelerate/vImageBoxConvolve_Planar8(_:_:_:_:_:_:_:_:_:)) Applies a box filter to an 8-bit planar source image. [`vImageBoxConvolve_ARGB8888`](/documentation/Accelerate/vImageBoxConvolve_ARGB8888(_:_:_:_:_:_:_:_:_:)) Applies a box filter to an 8-bit-per-channel, 4-channel interleaved source image. [`vImageTentConvolve_Planar8`](/documentation/Accelerate/vImageTentConvolve_Planar8(_:_:_:_:_:_:_:_:_:)) Applies a tent filter to an 8-bit planar source image. [`vImageTentConvolve_ARGB8888`](/documentation/Accelerate/vImageTentConvolve_ARGB8888(_:_:_:_:_:_:_:_:_:)) Applies a tent filter to an 8-bit-per-channel, 4-channel interleaved source image. ### Deconvolving [`vImageRichardsonLucyDeConvolve_Planar8`](/documentation/Accelerate/vImageRichardsonLucyDeConvolve_Planar8(_:_:_:_:_:_:_:_:_:_:_:_:_:_:_:_:)) Deconvolves an 8-bit planar image. [`vImageRichardsonLucyDeConvolve_PlanarF`](/documentation/Accelerate/vImageRichardsonLucyDeConvolve_PlanarF(_:_:_:_:_:_:_:_:_:_:_:_:_:_:)) Deconvolves a floating-point 32-bit planar image. [`vImageRichardsonLucyDeConvolve_ARGB8888`](/documentation/Accelerate/vImageRichardsonLucyDeConvolve_ARGB8888(_:_:_:_:_:_:_:_:_:_:_:_:_:_:_:_:)) Deconvolves an 8-bit-per-channel, 4-channel interleaved image. [`vImageRichardsonLucyDeConvolve_ARGBFFFF`](/documentation/Accelerate/vImageRichardsonLucyDeConvolve_ARGBFFFF(_:_:_:_:_:_:_:_:_:_:_:_:_:_:)) Deconvolves a floating-point 32-bit-per-channel, 4-channel interleaved image. ### Getting the minimum buffer size Use the [`kvImageGetTempBufferSize`](/documentation/Accelerate/kvImageGetTempBufferSize) A flag that returns the minimum temporary buffer size for the operation, given the parameters provided. flag with the appropriate convolution function. --- Copyright © 2026 Apple Inc. All rights reserved. | [Terms of Use](https://www.apple.com/legal/internet-services/terms/site.html) | [Privacy Policy](https://www.apple.com/privacy/privacy-policy)