RobustKernelImpl.h File Reference

(b9e049c (Sun Dec 31 11:36:26 2023 -0800))

#include <cmath>
#include "open3d/core/CUDAUtils.h"
#include "open3d/t/geometry/kernel/GeometryMacros.h"
#include "open3d/t/pipelines/registration/RobustKernel.h"

Go to the source code of this file.

Macros
#define	DISPATCH_ROBUST_KERNEL_FUNCTION(METHOD, scalar_t, scaling_parameter, shape_parameter, ...)
#define	DISPATCH_DUAL_ROBUST_KERNEL_FUNCTION(scalar_t, METHOD_1, scaling_parameter_1, METHOD_2, scaling_parameter_2, ...)

Macro Definition Documentation

DISPATCH_DUAL_ROBUST_KERNEL_FUNCTION

#define DISPATCH_DUAL_ROBUST_KERNEL_FUNCTION	(		scalar_t,
			METHOD_1,
			scaling_parameter_1,
			METHOD_2,
			scaling_parameter_2,
			...
	)

Parameters

scalar_t	type: float / double.
METHOD_1	registration::RobustKernelMethod Loss type.
scaling_parameter_1	Scaling parameter for loss fine-tuning.
METHOD_2	registration::RobustKernelMethod Loss type.
scaling_parameter_2	Scaling parameter for loss fine-tuning.

DISPATCH_ROBUST_KERNEL_FUNCTION

#define DISPATCH_ROBUST_KERNEL_FUNCTION	(		METHOD,
			scalar_t,
			scaling_parameter,
			shape_parameter,
			...
	)

To use Robust Kernel functions please refer the unit-tests for t::registration or the implementation use cases at t::pipelines::kernel::ComputePosePointToPlaneCUDA and t::pipelines::kernel::ComputePosePointToPlaneCPU.

Parameters

METHOD	registration::RobustKernelMethod Loss type.
scalar_t	type: float / double.
scaling_parameter	Scaling parameter for loss fine-tuning.
shape_parameter	Shape parameter for Generalized Loss method.