boofcv.alg.segmentation

Class ImageSegmentationOps

java.lang.Object

boofcv.alg.segmentation.ImageSegmentationOps

public class ImageSegmentationOps
extends java.lang.Object

Useful functions related to image segmentation

Constructor Summary

Constructors

Constructor and Description
ImageSegmentationOps()

Method Summary

Methods

Modifier and Type	Method and Description
static int	countRegionPixels(boofcv.struct.image.ImageSInt32 labeled, int which) Counts the number of instances of 'which' inside the labeled image.
static void	countRegionPixels(boofcv.struct.image.ImageSInt32 labeled, int totalRegions, int[] counts) Counts the number of pixels in all regions.
static void	markRegionBorders(boofcv.struct.image.ImageSInt32 labeled, boofcv.struct.image.ImageUInt8 output) Indicates border pixels between two regions.
static void	regionPixelId_to_Compact(boofcv.struct.image.ImageSInt32 graph, org.ddogleg.struct.GrowQueue_I32 segmentId, boofcv.struct.image.ImageSInt32 output) Compacts the region labels such that they are consecutive numbers starting from 0.

Methods inherited from class java.lang.Object

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

Constructor Detail

ImageSegmentationOps

public ImageSegmentationOps()

Method Detail

countRegionPixels

public static int countRegionPixels(boofcv.struct.image.ImageSInt32 labeled,
                    int which)

Counts the number of instances of 'which' inside the labeled image.

Parameters:

labeled - Image which has been labeled

which - The label being searched for

Returns:

Number of instances of 'which' in 'labeled'

countRegionPixels

public static void countRegionPixels(boofcv.struct.image.ImageSInt32 labeled,
                     int totalRegions,
                     int[] counts)

Counts the number of pixels in all regions. Regions must be have labels from 0 to totalRegions-1.

Parameters:

labeled - (Input) labeled image

totalRegions - Total number of regions

counts - Storage for pixel counts

regionPixelId_to_Compact

public static void regionPixelId_to_Compact(boofcv.struct.image.ImageSInt32 graph,
                            org.ddogleg.struct.GrowQueue_I32 segmentId,
                            boofcv.struct.image.ImageSInt32 output)

Compacts the region labels such that they are consecutive numbers starting from 0. The ID of a root node must the index of a pixel in the 'graph' image, taking in account the change in coordinates for sub-images.

Parameters:

graph - Input segmented image where the ID's are not compacted

segmentId - List of segment ID's. See comment above about what ID's are acceptable.

output - The new image after it has been compacted

markRegionBorders

public static void markRegionBorders(boofcv.struct.image.ImageSInt32 labeled,
                     boofcv.struct.image.ImageUInt8 output)

Indicates border pixels between two regions. If two adjacent pixels (4-connect) are not from the same region then both pixels are marked as true (value of 1) in output image, all other pixels are false (0).

Parameters:

labeled - Input segmented image.

output - Output binary image. 1 for border pixels.