Euclid::SpinImage< Mesh > Class Template Reference

The spin image descriptor. More...

#include <SpinImage.h>

Public Types
using	Vector_3 = Vector_3_t< Mesh >
using	FT = FT_t< Mesh >

Public Member Functions
void	build (const Mesh &mesh, const std::vector< Vector_3 > *vnormals=nullptr, FT resolution=0.0)
	Build up the necessary computational components. More...
template<typename Derived >
void	compute (Eigen::ArrayBase< Derived > &spin_img, float bin_scale=1.0f, int image_width=16, float support_angle=90.0f)
	Compute the spin image descriptor for all vertices. More...

Public Attributes
const Mesh *	mesh = nullptr
	The mesh being processed.
ProPtr< const std::vector< Vector_3 > >	vnormals = nullptr
	The vertex normals.
FT	resolution = 0.0
	The mesh resolution.

Detailed Description

template<typename Mesh>
class Euclid::SpinImage< Mesh >

Spin image is an image-based local shape descriptor. For every point on a mesh, an image is generated by projecting points onto the image plane within a local support.

Reference

Johnson A E, Hebert M. Using spin images for efficient object recognition in cluttered 3D scenes[J]. IEEE Transactions on pattern analysis and machine intelligence, 1999, 21(5): 433-449.

Member Function Documentation

template<typename Mesh >

void Euclid::SpinImage< Mesh >::build	(	const Mesh &	mesh,
		const std::vector< Vector_3 > *	vnormals = nullptr,
		FT	resolution = 0.0
	)

Parameters

mesh	The target mesh.
vnormals	The precomputed vertex normals, default to nullptr and it'll be computed internally.
resolution	The mesh resolution, i.e. avg. edge length. Default to 0.0 and it'll be computed internally.

template<typename Mesh >

template<typename Derived >

void Euclid::SpinImage< Mesh >::compute	(	Eigen::ArrayBase< Derived > &	spin_img,
		float	bin_scale = 1.0f,
		int	image_width = 16,
		float	support_angle = 90.0f
	)

Parameters

spin_img	The output spin image for v.
bin_scale	Multiple of the mesh resolution, default to 1.0 which is equavalent to use the mesh resolution as bin size.
image_width	Number of rows and columns for the image, default to bin_size * image_width equals to the actual image size in mesh coordinates.
support_angle	Maximum support angle in degrees, default to 90.0 thus supporting vertices pointing in the same direction as v.

---

The documentation for this class was generated from the following files:

• SpinImage.h
• SpinImage.cpp