gkdtree

Class GKDTree

java.lang.Object

gkdtree.GKDTree

public class GKDTree
extends java.lang.Object

Implementation of the Gaussian kd-tree. Supports the initialization and queries as well as dispaly functionality. "complicated theories only seem elegant in theory"

Author:

DarkMatt

Field Summary

Fields

Modifier and Type	Field and Description
static int	MAX_LEVEL
static int	NUM_SAMPLES

Constructor Summary

Constructors

Constructor and Description
GKDTree(java.lang.String filename, int dimension, FileOctree data, float scale_R)

Method Summary

Methods

Modifier and Type	Method and Description
void	blur(float sigma) Blur is a Query at every leaf node and propagates the value of the leaf node to adjacent leaf nodes.
void	downsample(float sigma) For all data-points do a Gaussian query - save results to leaf nodes Query results in a list of points.
float[]	get_drawingStructure() Returns quads at the cut values of the kd-tree.
float[]	get_drawingStructureColor()
float[]	get_drawingStructureUV() Returns uv-coordinates since they are required in the current rendering process.
float[]	get_Leaves() Returns a position array of all leaf nodes.
float[]	get_LeavesColor()
void	initTree(float sigma) Starts the tree initialization method.
java.util.ArrayList<GKDNode.Result>	queryIntoTree(float[] pos, float sigma, int num_samples) Starts a Gaussian query at the leaf node.
void	scale(float sigma_s, float sigma_r) Sets the scale parameters for the data-set.
void	undoScale()
void	upsample(float sigma) For all data-points a Query is executed.

Methods inherited from class java.lang.Object

equals, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Field Detail

NUM_SAMPLES

public static final int NUM_SAMPLES

See Also:

Constant Field Values

MAX_LEVEL

public static final int MAX_LEVEL

See Also:

Constant Field Values

Constructor Detail

GKDTree

public GKDTree(java.lang.String filename,
       int dimension,
       FileOctree data,
       float scale_R)

Method Detail

initTree

public void initTree(float sigma)

Starts the tree initialization method.

Parameters:

sigma -

scale

public void scale(float sigma_s,
         float sigma_r)

Sets the scale parameters for the data-set.

Parameters:

sigma_s -

sigma_r -

undoScale

public void undoScale()

downsample

public void downsample(float sigma)

For all data-points do a Gaussian query - save results to leaf nodes Query results in a list of points. Color values are assigned to the points according to the query origin value.

Parameters:

sigma -

blur

public void blur(float sigma)

Blur is a Query at every leaf node and propagates the value of the leaf node to adjacent leaf nodes. This requires a second value at every leaf node, or a copy of the tree.

Parameters:

sigma -

upsample

public void upsample(float sigma)

For all data-points a Query is executed. The values are sliced to the data-points.

Parameters:

sigma -

queryIntoTree

public java.util.ArrayList<GKDNode.Result> queryIntoTree(float[] pos,
                                                float sigma,
                                                int num_samples)

Starts a Gaussian query at the leaf node. (For details see paper.)

Parameters:

pos -

sigma -

num_samples -

Returns:

ArrayList

get_drawingStructure

public float[] get_drawingStructure()

Returns quads at the cut values of the kd-tree. (still not working 100% correctly... its drawn or all nodes)

Returns:

float[]

get_drawingStructureUV

public float[] get_drawingStructureUV()

Returns uv-coordinates since they are required in the current rendering process.

Returns:

float[]

get_drawingStructureColor

public float[] get_drawingStructureColor()

get_Leaves

public float[] get_Leaves()

Returns a position array of all leaf nodes. Need to be scaled.. since they are in the array! This is correct: Centroid of the bounding box, not centroid of the leaf node!

Returns:

float[]

get_LeavesColor

public float[] get_LeavesColor()