org.neuroph.nnet.comp.layer

Class FeatureMapsLayer

java.lang.Object

org.neuroph.core.Layer

org.neuroph.nnet.comp.layer.FeatureMapsLayer

All Implemented Interfaces:

java.io.Serializable, java.lang.Iterable<Neuron>

Direct Known Subclasses:

ConvolutionalLayer, InputMapsLayer, PoolingLayer

public abstract class FeatureMapsLayer
extends Layer

This class represents an array of feature maps which are 2 dimensional layers (Layer2D instances) and it is base class for Convolution and Pooling layers, which are used in ConvolutionalNetwork

See Also:

ConvolutionalLayer, PoolingLayer, ConvolutionalNetwork, Serialized Form

Constructor Summary

Constructors

Constructor and Description
FeatureMapsLayer() Creates a new empty feature maps layer with specified kernel
FeatureMapsLayer(Dimension2D mapDimensions) Creates a new empty feature maps layer with specified kernel and feature map dimensions.
FeatureMapsLayer(Dimension2D kernelDimension, Dimension2D mapDimensions, int mapCount, NeuronProperties neuronProp) Creates new feature maps layer with specified kernel and feature maps.
FeatureMapsLayer(Dimension2D mapDimensions, int mapCount, NeuronProperties neuronProp)

Method Summary

Modifier and Type	Method and Description
void	addFeatureMap(FeatureMapLayer featureMap) Adds a feature map (2d layer) to this feature map layer
abstract void	connectMaps(FeatureMapLayer fromMap, FeatureMapLayer toMap) Creates connections between two feature maps.
FeatureMapLayer	getFeatureMap(int index) Returns feature map (Layer2D) at specified index
java.util.List<FeatureMapLayer>	getFeatureMaps()
Dimension2D	getMapDimensions() Returns dimensions of feature maps in this layer
Neuron	getNeuronAt(int x, int y, int mapIndex) Returns neuron instance at specified (x, y) position at specified feature map layer
int	getNeuronsCount() Returns total number of neurons in all feature maps
int	getNumberOfMaps() Returns number of feature maps in this layer

Methods inherited from class org.neuroph.core.Layer

addNeuron, addNeuron, calculate, getLabel, getNeuronAt, getNeurons, getParentNetwork, indexOf, initializeWeights, isEmpty, iterator, removeAllNeurons, removeNeuron, removeNeuronAt, reset, setLabel, setNeuron, setParentNetwork

Methods inherited from class java.lang.Object

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

Methods inherited from interface java.lang.Iterable

forEach, spliterator

Constructor Detail

FeatureMapsLayer

public FeatureMapsLayer()

Creates a new empty feature maps layer with specified kernel

Parameters:

kernel - kernel to use for all feature maps

FeatureMapsLayer

public FeatureMapsLayer(Dimension2D mapDimensions)

Creates a new empty feature maps layer with specified kernel and feature map dimensions.

Parameters:

kernel - kernel used for all feature maps in this layer

mapDimensions - mapDimensions of feature maps in this layer

FeatureMapsLayer

public FeatureMapsLayer(Dimension2D kernelDimension,
                        Dimension2D mapDimensions,
                        int mapCount,
                        NeuronProperties neuronProp)

Creates new feature maps layer with specified kernel and feature maps. Also creates feature maps and neurons in feature maps;

Parameters:

kernel - kernel used for all feature maps in this layer

mapDimensions - mapDimensions of feature maps in this layer

mapCount - number of feature maps

neuronProp - properties for neurons in feature maps

FeatureMapsLayer

public FeatureMapsLayer(Dimension2D mapDimensions,
                        int mapCount,
                        NeuronProperties neuronProp)

Method Detail

getFeatureMaps

public java.util.List<FeatureMapLayer> getFeatureMaps()

addFeatureMap

public void addFeatureMap(FeatureMapLayer featureMap)

Adds a feature map (2d layer) to this feature map layer

Parameters:

featureMap - feature map to add

getFeatureMap

public FeatureMapLayer getFeatureMap(int index)

Returns feature map (Layer2D) at specified index

Parameters:

index - index of feature map

Returns:

feature map (Layer2D instance) at specified index

getNumberOfMaps

public int getNumberOfMaps()

Returns number of feature maps in this layer

Returns:

number of feature maps in this layer

getNeuronAt

public Neuron getNeuronAt(int x,
                          int y,
                          int mapIndex)

Returns neuron instance at specified (x, y) position at specified feature map layer

Parameters:

x - neuron's x position

y - neuron's y position

mapIndex - feature map index

Returns:

neuron at specified (x, y, map) position

getNeuronsCount

public int getNeuronsCount()

Returns total number of neurons in all feature maps

Overrides:

getNeuronsCount in class Layer

Returns:

total number of neurons in all feature maps

getMapDimensions

public Dimension2D getMapDimensions()

Returns dimensions of feature maps in this layer

Returns:

dimensions of feature maps in this layer

connectMaps

public abstract void connectMaps(FeatureMapLayer fromMap,
                                 FeatureMapLayer toMap)

Creates connections between two feature maps. It does nothing here, connectivity patterns are defined by subclasses... Maybe it should be even removed from here or made abstract......

Parameters:

fromMap -

toMap -