ImFusion::Curve Class Reference

#include <ImFusion/Base/Curve.h>

Interpolation and fitting of a 1D curve. More...

Detailed Description

Interpolation and fitting of a 1D curve.

If no x values are provided, equal-spaced data locations within 0..1 are used. Interpolation is done with Hermite cubic splines and conditioning to avoid overshoot of the curve.

Public Member Functions
	Curve (const std::vector< double > &y)
	Creates curve with equal-spaced data values to interpolate.
	Curve (const std::vector< double > &x, const std::vector< double > &y)
	Creates curve with arbitrary data positions and values.
	Curve (const std::vector< vec3 > &points)
bool	fit (int n, bool adaptConst=false, bool optimize=false)
	Fit curve using desired dimension, optionally optimizing point locations.
const std::vector< double > &	x () const
	Access to the support points of the curve.
const std::vector< double > &	y () const
	Access to the function values of the curve.

Curve and error evaluation
Note Perform an evaluation or error computation only after a successful run of fit(). Otherwise invalid results are returned.
double	evaluate (double x) const
std::vector< double >	evaluate (const std::vector< double > &x) const
	Evaluate the curve function at multiple locations.
void	evaluate (std::vector< vec3 > &points) const
	Evaluate the curve at multiple locations and write the result back in the same vector.
std::vector< double >	evaluate (int n) const
	Evaluate the curve with a chosen resolution.
double	error (const std::vector< double > &y) const
	Compute the mean error of the curve and the provided regular spaced data.
double	error (const std::vector< double > &x, const std::vector< double > &y) const
	Compute the mean error of the curve and the provided data in x and y.

Member Function Documentation

fit()

bool ImFusion::Curve::fit	(	int	n,
		bool	adaptConst = false,
		bool	optimize = false )

Fit curve using desired dimension, optionally optimizing point locations.

Returns

False if no data point locations given in m_x, true otherwise.

evaluate() [1/2]

double ImFusion::Curve::evaluate

(

double

x

)

const

Evaluate the curve function at a single location

evaluate() [2/2]

std::vector< double > ImFusion::Curve::evaluate

(

int

n

)

const

Evaluate the curve with a chosen resolution.

The curve is evaluated with n samples in a regular grid defined by the minimum and maximum of the data point locations (not necessarily 0..1).

---

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

• ImFusion/Base/Curve.h