BpOptimizeCUDA Directory Reference

Directory dependency graph for BpOptimizeCUDA:

Files
file	KernelBpStereo.cu [code]
	This file defines the methods to perform belief propagation for disparity map estimation from stereo images on CUDA.
file	KernelBpStereoDebug.cu [code]
	This file defines CUDA kernel functions for debugging belief propagation processing.
file	KernelBpStereoHalf.cu [code]
	This file defines the template specialization to perform belief propagation using half precision for disparity map estimation from stereo images on CUDA to prevent overflow in val_to_normalize message value computation shouldn't be an issue if using bfloat instead of half so recommend using that instead of using these template specialization functions.
file	KernelFilter.cu [code]
	This kernel is used to filter the image with the given filter in the vertical and horizontal directions.
file	KernelFilterHeader.cuh [code]
	Header for the kernel to apply a horizontal/vertical filter to image data.
file	ProcessBpCUDA.cpp [code]
	Defines the functions to run the CUDA implementation of 2D Stereo estimation using BP.
file	ProcessBpCUDA.h [code]
file	RunBpOnStereoSetCUDA.cpp [code]
	Defines the methods to run BP Stereo implementation on a series of images using various options.
file	RunBpOnStereoSetCUDA.h [code]
	Declares child class of RunBpOnStereoSet to run CUDA implementation of belief propagation on a given stereo set as defined by reference and test image file paths.
file	SmoothImageCUDA.cpp [code]
file	SmoothImageCUDA.h [code]
	Declares child class of SmoothImage for smoothing images in the CUDA implementation.