ImageDev

CudaBoxFilter2d

Smoothes an image with a box kernel. The calculations are performed on the GPU.

Access to parameter description

This command is experimental, his signature may be modified between now and his final version.

 For an introduction to image filters: section Images Filtering.

This algorithm smoothes an image using the same box kernel as a lowpass filter. The filter's X and Y sizes are user-defined.

The algorithm calculates the local mean in a given size window. For a window of size $2p+1$ in X and $2q+1$ in Y: $$ O(n,m)=\frac{1}{K}\sum_{i=-p}^{p}\sum_{j=-q}^{q} I(n-i,m-j) $$ The $K$ coefficient is a normalization factor: Note: If the result is normalized, the output image has same type as the input. Else, it is upgraded according to the Image type promotion rule.

See also

Function Syntax

This function returns outputImage. 
// Function prototype
std::shared_ptr< iolink::ImageView >
cudaBoxFilter2d( std::shared_ptr< iolink::ImageView > inputImage,
                 uint32_t kernelSizeX,
                 uint32_t kernelSizeY,
                 CudaBoxFilter2d::TilingMode tilingMode,
                 iolink::Vector2u32 tileSize,
                 CudaContext::Ptr cudaContext,
                 std::shared_ptr< iolink::ImageView > outputImage = nullptr );
This function returns outputImage. 
// Function prototype.
cuda_box_filter_2d(input_image: idt.ImageType,
                   kernel_size_x: int = 3,
                   kernel_size_y: int = 3,
                   tiling_mode: CudaBoxFilter2d.TilingMode = CudaBoxFilter2d.TilingMode.NONE,
                   tile_size: Iterable[int] = [1024, 1024],
                   cuda_context: Union[CudaContext, None] = None,
                   output_image: idt.ImageType = None) -> idt.ImageType
This function returns outputImage. 
// Function prototype.
public static IOLink.ImageView
CudaBoxFilter2d( IOLink.ImageView inputImage,
                 UInt32 kernelSizeX = 3,
                 UInt32 kernelSizeY = 3,
                 CudaBoxFilter2d.TilingMode tilingMode = ImageDev.CudaBoxFilter2d.TilingMode.NONE,
                 uint[] tileSize = null,
                 Data.CudaContext cudaContext = null,
                 IOLink.ImageView outputImage = null );

Class Syntax

Parameters

Parameter Name Description Type Supported Values Default Value
input
inputImage
The input image. Image Binary, Grayscale or Multispectral nullptr
input
kernelSizeX
The horizontal kernel size. UInt32 >=1 3
input
kernelSizeY
The vertical kernel size. UInt32 >=1 3
input
tilingMode
The way to manage the GPU memory.
NONE The entire input image is transferred to the GPU memory. If the total input, intermediate and output data size exceed the GPU memory, the computation will fail.
USER_DEFINED The input image is processed by tiles of a predefined size.
Enumeration NONE
input
tileSize
The tile width and height in pixels. This parameter is used only in USER_DEFINED tiling mode. Vector2u32 >=1 {1024, 1024}
input
cudaContext
CUDA context information. CudaContext nullptr
output
outputImage
The output image. Its dimensions are forced to the same values as the input. Its type is the same as the input if the normalization is set to yes, else the type is upgraded. Image nullptr
Parameter Name Description Type Supported Values Default Value
input
input_image
The input image. image Binary, Grayscale or Multispectral None
input
kernel_size_x
The horizontal kernel size. uint32 >=1 3
input
kernel_size_y
The vertical kernel size. uint32 >=1 3
input
tiling_mode
The way to manage the GPU memory.
NONE The entire input image is transferred to the GPU memory. If the total input, intermediate and output data size exceed the GPU memory, the computation will fail.
USER_DEFINED The input image is processed by tiles of a predefined size.
enumeration NONE
input
tile_size
The tile width and height in pixels. This parameter is used only in USER_DEFINED tiling mode. vector2u32 >=1 [1024, 1024]
input
cuda_context
CUDA context information. cuda_context None
output
output_image
The output image. Its dimensions are forced to the same values as the input. Its type is the same as the input if the normalization is set to yes, else the type is upgraded. image None
Parameter Name Description Type Supported Values Default Value
input
inputImage
The input image. Image Binary, Grayscale or Multispectral null
input
kernelSizeX
The horizontal kernel size. UInt32 >=1 3
input
kernelSizeY
The vertical kernel size. UInt32 >=1 3
input
tilingMode
The way to manage the GPU memory.
NONE The entire input image is transferred to the GPU memory. If the total input, intermediate and output data size exceed the GPU memory, the computation will fail.
USER_DEFINED The input image is processed by tiles of a predefined size.
Enumeration NONE
input
tileSize
The tile width and height in pixels. This parameter is used only in USER_DEFINED tiling mode. Vector2u32 >=1 {1024, 1024}
input
cudaContext
CUDA context information. CudaContext null
output
outputImage
The output image. Its dimensions are forced to the same values as the input. Its type is the same as the input if the normalization is set to yes, else the type is upgraded. Image null

Object Examples

auto polystyrene = ioformat::readImage( std::string( IMAGEDEVDATA_IMAGES_FOLDER ) + "polystyrene.tif" );

CudaBoxFilter2d cudaBoxFilter2dAlgo;
cudaBoxFilter2dAlgo.setInputImage( polystyrene );
cudaBoxFilter2dAlgo.setKernelSizeX( 3 );
cudaBoxFilter2dAlgo.setKernelSizeY( 3 );
cudaBoxFilter2dAlgo.setTilingMode( CudaBoxFilter2d::TilingMode::NONE );
cudaBoxFilter2dAlgo.setTileSize( {264, 264} );
cudaBoxFilter2dAlgo.setCudaContext( nullptr );
cudaBoxFilter2dAlgo.setOutputImage( iolink::ImageViewFactory::allocate( iolink::VectorXu64( { 1, 1 } ), iolink::DataTypeId::UINT8 ) );
cudaBoxFilter2dAlgo.execute();

std::cout << "outputImage:" << cudaBoxFilter2dAlgo.outputImage()->toString();

polystyrene = ioformat.read_image(imagedev_data.get_image_path("polystyrene.tif"))

cuda_box_filter_2d_algo = imagedev.CudaBoxFilter2d()
cuda_box_filter_2d_algo.input_image = polystyrene
cuda_box_filter_2d_algo.kernel_size_x = 3
cuda_box_filter_2d_algo.kernel_size_y = 3
cuda_box_filter_2d_algo.tiling_mode = imagedev.CudaBoxFilter2d.NONE
cuda_box_filter_2d_algo.tile_size = [264, 264]
cuda_box_filter_2d_algo.cuda_context = None
cuda_box_filter_2d_algo.execute()

print("output_image:", str(cuda_box_filter_2d_algo.output_image))

ImageView polystyrene = ViewIO.ReadImage( @"Data/images/polystyrene.tif" );

CudaBoxFilter2d cudaBoxFilter2dAlgo = new CudaBoxFilter2d
{
    inputImage = polystyrene,
    kernelSizeX = 3,
    kernelSizeY = 3,
    tilingMode = CudaBoxFilter2d.TilingMode.NONE,
    tileSize = new uint[]{264, 264},
    cudaContext = null
};
cudaBoxFilter2dAlgo.Execute();

Console.WriteLine( "outputImage:" + cudaBoxFilter2dAlgo.outputImage.ToString() );

Function Examples

auto polystyrene = ioformat::readImage( std::string( IMAGEDEVDATA_IMAGES_FOLDER ) + "polystyrene.tif" );

auto result = cudaBoxFilter2d( polystyrene, 3, 3, CudaBoxFilter2d::TilingMode::NONE, {264, 264}, nullptr , iolink::ImageViewFactory::allocate( iolink::VectorXu64( { 1, 1 } ), iolink::DataTypeId::UINT8 ));

std::cout << "outputImage:" << result->toString();

polystyrene = ioformat.read_image(imagedev_data.get_image_path("polystyrene.tif"))

result = imagedev.cuda_box_filter_2d(polystyrene, 3, 3, imagedev.CudaBoxFilter2d.NONE, [264, 264], None)

print("output_image:", str(result))

ImageView polystyrene = ViewIO.ReadImage( @"Data/images/polystyrene.tif" );

IOLink.ImageView result = Processing.CudaBoxFilter2d( polystyrene, 3, 3, CudaBoxFilter2d.TilingMode.NONE, new uint[]{264, 264}, null );

Console.WriteLine( "outputImage:" + result.ToString() );