Package qupath.lib.classifiers.pixel

Interface PixelClassifier

public interface PixelClassifier

Interface defining a pixel classifier.

Pixel classifiers can be applied directly to an image, typically using colors and textures.

Field Summary

Fields

Modifier and Type	Field	Description
static final String	PROJECT_LOCATION	Default location to use when storing object classifiers in a project.

Method Summary

Modifier and Type	Method	Description
BufferedImage	applyClassification(ImageData<BufferedImage> server, RegionRequest request)	Apply pixel classifier to a specified region of an image.
PixelClassifierMetadata	getMetadata()	Get metadata that describes how the classifier should be called, and the kind of output it provides.
boolean	supportsImage(ImageData<BufferedImage> imageData)	Query whether the classifier supports a particular image.

Field Details

PROJECT_LOCATION

static final String PROJECT_LOCATION

Default location to use when storing object classifiers in a project.

See Also:

• Project.getResources(String, Class, String)
• Constant Field Values

Method Details

supportsImage

boolean supportsImage(ImageData<BufferedImage> imageData)

Query whether the classifier supports a particular image. It may not if the number of channels (for example) is incompatible.

Parameters:

imageData -

Returns:

applyClassification

BufferedImage applyClassification(ImageData<BufferedImage> server,
 RegionRequest request)
                           throws IOException

Apply pixel classifier to a specified region of an image.

An ImageData and RegionRequest are supplied, rather than a BufferedImage directly, because there may be a need to adapt to the image resolution and/or incorporate padding to reduce boundary effects.

There is no guarantee that the returned BufferedImage will be the same size as the input region (after downsampling), but rather that it should contain the full classification information for the specified region.

Practically, this means that there may be fewer pixels in the output because the classification inherently involves downsampling.

Parameters:

server -

request -

Returns:

a BufferedImage representing the pixel classifications as separate bands.

Throws:

IOException - if unable to read pixels from server

getMetadata

PixelClassifierMetadata getMetadata()

Get metadata that describes how the classifier should be called, and the kind of output it provides.

Returns: