In This Topic ▼

LBitmap::KMeansBitmapSegmentation

#include "ltwrappr.h"

virtual L_INT LBitmap::KMeansBitmapSegmentation(nCluster, ppOutCenters, nOutCentersCount, pInCenters, uFlags)

L_UINT nCluster;	an integer that represents the number of output clusters or colors in the output image
L_COLORREF** ppOutCenters;	pointer of a pointer to L_COLORREF values that will be filled with the centers of the output clusters
L_INT* nOutCentersCount;	pointer to an Integer that represents the number of output clusters
L_COLORREF* pInCenters;	pointer to L_COLORREF values that represents the initial centers of the clusters. Used whenever the Type is KMEANS_USERDEFINED input clusters
L_UINT uFlags;	flag that determines which initializing algorithm to use when choosing the initial centers for the clusters

Performs image segmentation and color reduction using the K-means algorithm.

Parameter	Description
nCluster	An integer that represents the number of output clusters or colors in the output image. Valid values range from 2 to 255. The default value is 4.
ppOutCenters	Pointer of a pointer to L_COLORREF values that will be filled with the centers of the output clusters.
nOutCentersCount	Pointer to an integer that represents the number of output clusters.
pInCenters	Pointer to L_COLORREF values that represents the initial centers of the clusters. Used whenever the Type is KMEANS_USERDEFINED input clusters.
uFlags	Flag that specifies which initializing algorithm to use when choosing the initial centers for the clusters.
	Value	Meaning
	KMEANS_RANDOM	[0x00000002] Use a Random sampling algorithm to set the initial means. This is the default value.
	KMEANS_UNIFORM	[0x00000003] Use a Uniform sampling algorithm to set the initial means.
	KMEANS_USERDEFINED	[0x00000004] Use an array of user-defined initial centers to set the initial means.

Returns

SUCCESS	The function was successful.
< 1	An error occurred. Refer to Return Codes.

Comments

K-Means is an algorithm for analyzing data. Each observation gets placed in the cluster having the mean nearest it. The number of clusters returned are less than or equal to the number of input clusters. If the image contains fewer clusters than the number of input clusters, the ppOutCenters parameter will be filled with the centers of the output clusters.

This function can only process entire images. It does not support regions.

This function supports 12- and 16-bit grayscale and 48- and 64-bit color images.

This function supports signed/unsigned images.

Required DLLs and Libraries

LTDIS
LTFIL
LTIMGCOR

For a listing of the exact DLLs and Libraries needed, based on the toolkit version, refer to Files To Be Included With Your Application.

Platforms

Win32, x64.

Example

#if defined (LEADTOOLS_V19_OR_LATER) 
L_INT LBitmap__KMeansBitmapSegmentationExample(L_VOID) 
{ 
    L_INT nRet; 
    LBitmap LeadBitmap; 
 
    nRet = LeadBitmap.Load(MAKE_IMAGE_PATH(TEXT("IMAGE3.dcm")), 0,ORDER_BGR); 
    if(nRet !=SUCCESS) 
        return nRet ; 
 
    L_COLORREF* poutCenters = NULL ; 
    L_INT outCount ; 
 
    nRet = LeadBitmap.KMeansBitmapSegmentation( 5, 
                                                &poutCenters, 
                                                &outCount, 
                                                NULL, 
                                                KMEANS_RANDOM) ; 
 
    LeadBitmap.FreeKmeansOutput(poutCenters); 
 
    return nRet ; 
} 
#endif // LEADTOOLS_V19_OR_LATER

Download our FREE evaluation

Help Version 20.0.2020.4.5

LEADTOOLS Raster Imaging C++ Class Library Help

Introduction

Version History

Getting Started

LEADTOOLS Raster Imaging Features

Redistributables/Files To Be Included With Your Application

Quick Reference

Tutorials

Classes and Member Functions

Summary of Features

Functions:	LBitmap::LevelsetBitmapRgn, LBitmap::LambdaConnectedness, LBitmap::ShrinkWrapTool, LBitmap::WatershedBitmap, LBitmap::KMeansBitmapSegmentation, LBitmap::GWireGetMinPath, Class Members
	Raster Image Functions: Image Analysis
	Processing an Image