org.openimaj.image.processing.face.feature.LocalLBPHistogram Maven / Gradle / Ivy
Go to download
Show more of this group Show more artifacts with this name
Show all versions of faces Show documentation
Show all versions of faces Show documentation
Implementation of a flexible face-recognition pipeline,
including pluggable detectors, aligners, feature extractors
and recognisers.
/**
* Copyright (c) 2011, The University of Southampton and the individual contributors.
* All rights reserved.
*
* Redistribution and use in source and binary forms, with or without modification,
* are permitted provided that the following conditions are met:
*
* * Redistributions of source code must retain the above copyright notice,
* this list of conditions and the following disclaimer.
*
* * Redistributions in binary form must reproduce the above copyright notice,
* this list of conditions and the following disclaimer in the documentation
* and/or other materials provided with the distribution.
*
* * Neither the name of the University of Southampton nor the names of its
* contributors may be used to endorse or promote products derived from this
* software without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND
* ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
* WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
* DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR
* ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
* (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
* LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON
* ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
* (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
* SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
package org.openimaj.image.processing.face.feature;
import java.io.DataInput;
import java.io.DataOutput;
import java.io.IOException;
import org.openimaj.feature.FeatureVectorProvider;
import org.openimaj.feature.FloatFV;
import org.openimaj.image.FImage;
import org.openimaj.image.feature.dense.binarypattern.ExtendedLocalBinaryPattern;
import org.openimaj.image.feature.dense.binarypattern.UniformBinaryPattern;
import org.openimaj.image.processing.face.alignment.FaceAligner;
import org.openimaj.image.processing.face.alignment.IdentityAligner;
import org.openimaj.image.processing.face.detection.DetectedFace;
import org.openimaj.io.IOUtils;
/**
* A {@link FacialFeature} built from decomposing the face image into
* (non-overlapping) blocks and building histograms of the
* {@link ExtendedLocalBinaryPattern}s for each block and then concatenating to
* form the final feature.
*
* @author Jonathon Hare ([email protected])
*/
public class LocalLBPHistogram implements FacialFeature, FeatureVectorProvider {
/**
* A {@link FacialFeatureExtractor} for building {@link LocalLBPHistogram}s.
*
* @author Jonathon Hare ([email protected])
*
* @param
* Type of {@link DetectedFace}.
*/
public static class Extractor implements FacialFeatureExtractor {
FaceAligner aligner;
int blocksX = 25;
int blocksY = 25;
int samples = 8;
int radius = 1;
/**
* Construct with a {@link IdentityAligner}
*/
public Extractor() {
this.aligner = new IdentityAligner();
}
/**
* Construct with the given aligner.
*
* @param aligner
* the aligner
*/
public Extractor(FaceAligner aligner) {
this.aligner = aligner;
}
/**
* Construct with the given aligner, parameters describing how the image
* is broken into blocks and parameters describing the radius of the LBP
* extraction circle, and how many samples are made.
*
* @param aligner
* The face aligner
* @param blocksX
* The number of blocks in the x-direction
* @param blocksY
* The number of blocks in the y-direction
* @param samples
* The number of samples around the circle for the
* {@link ExtendedLocalBinaryPattern}
* @param radius
* the radius used for the {@link ExtendedLocalBinaryPattern}
* .
*/
public Extractor(FaceAligner aligner, int blocksX, int blocksY, int samples, int radius) {
this.aligner = aligner;
this.blocksX = blocksX;
this.blocksY = blocksY;
this.samples = samples;
this.radius = radius;
}
@Override
public LocalLBPHistogram extractFeature(T detectedFace) {
final LocalLBPHistogram f = new LocalLBPHistogram();
final FImage face = aligner.align(detectedFace);
final FImage mask = aligner.getMask();
f.initialise(face, mask, blocksX, blocksY, samples, radius);
return f;
}
@Override
public void readBinary(DataInput in) throws IOException {
final String alignerClass = in.readUTF();
aligner = IOUtils.newInstance(alignerClass);
aligner.readBinary(in);
blocksX = in.readInt();
blocksY = in.readInt();
radius = in.readInt();
samples = in.readInt();
}
@Override
public byte[] binaryHeader() {
return this.getClass().getName().getBytes();
}
@Override
public void writeBinary(DataOutput out) throws IOException {
out.writeUTF(aligner.getClass().getName());
aligner.writeBinary(out);
out.writeInt(blocksX);
out.writeInt(blocksY);
out.writeInt(radius);
out.writeInt(samples);
}
@Override
public String toString() {
return String.format("LocalLBPHistogram.Factory[blocksX=%d,blocksY=%d,samples=%d,radius=%d]", blocksX,
blocksY, samples, radius);
}
}
float[][][] histograms;
transient FloatFV featureVector;
protected void initialise(FImage face, FImage mask, int blocksX, int blocksY, int samples, int radius) {
final int[][] pattern = ExtendedLocalBinaryPattern.calculateLBP(face, radius, samples);
final boolean[][][] maps = UniformBinaryPattern.extractPatternMaps(pattern, samples);
final int bx = face.width / blocksX;
final int by = face.height / blocksY;
histograms = new float[blocksY][blocksX][maps.length];
// build histogram
for (int p = 0; p < maps.length; p++) {
for (int y = 0; y < blocksY; y++) {
for (int x = 0; x < blocksX; x++) {
for (int j = 0; j < by; j++) {
for (int i = 0; i < bx; i++) {
if (maps[p][y * by + j][x * bx + i])
histograms[y][x][p]++;
}
}
}
}
}
// normalise
for (int y = 0; y < blocksY; y++) {
for (int x = 0; x < blocksX; x++) {
float count = 0;
for (int p = 0; p < maps.length; p++) {
count += histograms[y][x][p];
}
for (int p = 0; p < maps.length; p++) {
histograms[y][x][p] /= count;
}
}
}
updateFeatureVector();
}
protected void updateFeatureVector() {
featureVector = new FloatFV(histograms.length * histograms[0].length * histograms[0][0].length);
int i = 0;
for (int y = 0; y < histograms.length; y++) {
for (int x = 0; x < histograms[0].length; x++) {
for (int p = 0; p < histograms[0][0].length; p++) {
featureVector.values[i] = histograms[y][x][p];
i++;
}
}
}
}
@Override
public byte[] binaryHeader() {
return "LBPH".getBytes();
}
@Override
public void readBinary(DataInput in) throws IOException {
final int by = in.readInt();
final int bx = in.readInt();
final int p = in.readInt();
histograms = new float[by][bx][p];
for (int j = 0; j < by; j++) {
for (int i = 0; i < bx; i++) {
for (int k = 0; k < p; k++) {
histograms[j][i][k] = in.readFloat();
}
}
}
updateFeatureVector();
}
@Override
public void writeBinary(DataOutput out) throws IOException {
out.writeInt(histograms.length);
out.writeInt(histograms[0].length);
out.writeInt(histograms[0][0].length);
for (final float[][] hist1 : histograms) {
for (final float[] hist2 : hist1) {
for (final float h : hist2) {
out.writeFloat(h);
}
}
}
}
@Override
public FloatFV getFeatureVector() {
if (featureVector == null)
updateFeatureVector();
return featureVector;
}
}