Optimized Belief Propagation (CPU and GPU)
|
Header for the kernel to apply a horizontal/vertical filter to image data. More...
#include <cuda_runtime.h>
#include <type_traits>
Go to the source code of this file.
Namespaces | |
beliefprop_cuda | |
Namespace to define global kernel functions for parallel belief propagation processing using CUDA. | |
Macros | |
#define | KERNEL_FILTER_HEADER_CUH |
Functions | |
__global__ void | beliefprop_cuda::convertUnsignedIntImageToFloat (const unsigned int *uint_image_pixels, float *float_image_pixels, unsigned int width_images, unsigned int height_images) |
Kernel to convert the unsigned int pixels to float pixels in an image when smoothing is not desired but the pixels need to be converted to floats the input image is stored as unsigned ints output filtered image stored in float_image_pixels. More... | |
template<BpImData_t T> | |
__global__ void | beliefprop_cuda::FilterImageAcross (const T *image_to_filter, float *filtered_image, unsigned int width_images, unsigned int height_images, const float *image_filter, unsigned int size_filter) |
Kernel to apply a horizontal filter on each pixel of the image in parallel the input image is stored as unsigned ints the output filtered image is returned as an array of floats. More... | |
template<BpImData_t T> | |
__global__ void | beliefprop_cuda::FilterImageVertical (const T *image_to_filter, float *filtered_image, unsigned int width_images, unsigned int height_images, const float *image_filter, unsigned int size_filter) |
Kernel to apply a vertical filter on each pixel of the image in parallel the input image is stored as unsigned ints the output filtered image is returned as an array of floats. More... | |
Header for the kernel to apply a horizontal/vertical filter to image data.
Definition in file KernelFilterHeader.cuh.
#define KERNEL_FILTER_HEADER_CUH |
Definition at line 28 of file KernelFilterHeader.cuh.