org.openimaj.demos.FVFWExperiment Maven / Gradle / Ivy
Go to download
Show more of this group Show more artifacts with this name
Show all versions of sandbox Show documentation
Show all versions of sandbox Show documentation
A project for various tests that don't quite constitute
demos but might be useful to look at.
/**
* 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.demos;
import java.io.File;
import java.io.FileNotFoundException;
import java.io.IOException;
import java.util.ArrayList;
import java.util.List;
import java.util.Random;
import java.util.Scanner;
import org.openimaj.feature.FloatFV;
import org.openimaj.io.IOUtils;
import org.openimaj.ml.linear.projection.LargeMarginDimensionalityReduction;
import Jama.Matrix;
import com.jmatio.io.MatFileReader;
import com.jmatio.types.MLSingle;
public class FVFWExperiment {
// private static final String FOLDER =
// "lfw-centre-affine-pdsift-pca64-augm-fv512/";
// private static final String FOLDER = "lfw-centre-affine-matlab-fisher/";
private static final String FOLDER = "matlab-fvs/";
static class FacePair {
boolean same;
File firstFV;
File secondFV;
public FacePair(File first, File second, boolean same) {
this.firstFV = first;
this.secondFV = second;
this.same = same;
}
FloatFV loadFirst() throws IOException {
return IOUtils.read(firstFV, FloatFV.class);
}
FloatFV loadSecond() throws IOException {
return IOUtils.read(secondFV, FloatFV.class);
}
}
static class Subset {
List testPairs = new ArrayList();
List trainingPairs = new ArrayList();
}
static List loadSubsets() throws IOException {
final List subsets = new ArrayList();
for (int i = 0; i < 10; i++)
subsets.add(new Subset());
loadPairs(new File("/Users/jon/Data/lfw/pairs.txt"), subsets);
loadPeople(new File("/Users/jon/Data/lfw/people.txt"), subsets);
return subsets;
}
private static void loadPairs(File file, List subsets) throws FileNotFoundException {
final Scanner sc = new Scanner(file);
final int nsets = sc.nextInt();
final int nhpairs = sc.nextInt();
if (nsets != 10 || nhpairs != 300) {
sc.close();
throw new RuntimeException();
}
for (int s = 0; s < 10; s++) {
for (int i = 0; i < 300; i++) {
final String name = sc.next();
final int firstIdx = sc.nextInt();
final int secondIdx = sc.nextInt();
final File first = new File(file.getParentFile(), FOLDER + name
+ "/" + name + String.format("_%04d.bin", firstIdx));
final File second = new File(file.getParentFile(), FOLDER + name
+ "/" + name + String.format("_%04d.bin", secondIdx));
subsets.get(s).testPairs.add(new FacePair(first, second, true));
}
for (int i = 0; i < 300; i++) {
final String firstName = sc.next();
final int firstIdx = sc.nextInt();
final String secondName = sc.next();
final int secondIdx = sc.nextInt();
final File first = new File(file.getParentFile(), FOLDER
+ firstName
+ "/" + firstName + String.format("_%04d.bin", firstIdx));
final File second = new File(file.getParentFile(), FOLDER
+ secondName
+ "/" + secondName + String.format("_%04d.bin", secondIdx));
subsets.get(s).testPairs.add(new FacePair(first, second, false));
}
}
sc.close();
}
private static void loadPeople(File file, List subsets) throws FileNotFoundException {
final Scanner sc = new Scanner(file);
final int nsets = sc.nextInt();
if (nsets != 10) {
sc.close();
throw new RuntimeException();
}
for (int s = 0; s < 10; s++) {
final int nnames = sc.nextInt();
final List files = new ArrayList(nnames);
for (int i = 0; i < nnames; i++) {
final String name = sc.next();
final int numPeople = sc.nextInt();
for (int j = 1; j <= numPeople; j++) {
final File f = new File(file.getParentFile(), FOLDER + name
+ "/" + name + String.format("_%04d.bin", j));
files.add(f);
}
}
for (int i = 0; i < files.size(); i++) {
final File first = files.get(i);
for (int j = i + 1; j < files.size(); j++) {
final File second = files.get(j);
final boolean same = first.getName().substring(0, first.getName().lastIndexOf("_"))
.equals(second.getName().substring(0, second.getName().lastIndexOf("_")));
subsets.get(s).trainingPairs.add(new FacePair(first, second, same));
subsets.get(s).trainingPairs.add(new FacePair(second, first, same));
}
}
}
sc.close();
}
static Subset createExperimentalFold(List subsets, int foldIdx) {
final Subset subset = new Subset();
// testing data is from the indexed fold
subset.testPairs = subsets.get(foldIdx).testPairs;
// training data is from the other folds
final List training = new ArrayList();
for (int i = 0; i < foldIdx; i++)
training.addAll(subsets.get(i).trainingPairs);
for (int i = foldIdx + 1; i < subsets.size(); i++)
training.addAll(subsets.get(i).trainingPairs);
subset.trainingPairs = reorder(training);
return subset;
}
private static List reorder(List training) {
final List trainingTrue = new ArrayList();
final List trainingFalse = new ArrayList();
for (final FacePair fp : training) {
if (fp.same)
trainingTrue.add(fp);
else
trainingFalse.add(fp);
}
resample(trainingTrue, 4000000);
resample(trainingFalse, 4000000);
final List trainingResorted = new ArrayList();
for (int i = 0; i < trainingTrue.size(); i++) {
trainingResorted.add(trainingTrue.get(i));
trainingResorted.add(trainingFalse.get(i));
}
return trainingResorted;
}
private static void resample(List pairs, int sz) {
final List oldPairs = new ArrayList(sz);
oldPairs.addAll(pairs);
pairs.clear();
final Random r = new Random();
for (int i = 0; i < sz; i++) {
pairs.add(oldPairs.get(r.nextInt(oldPairs.size())));
}
}
public static void main(String[] args) throws IOException {
final List subsets = loadSubsets();
final Subset fold = createExperimentalFold(subsets, 1);
// // final LargeMarginDimensionalityReduction lmdr = new
// // LargeMarginDimensionalityReduction(128);
// final LargeMarginDimensionalityReduction lmdr = loadMatlabPCAW();
//
// final double[][] fInit = new double[1000][];
// final double[][] sInit = new double[1000][];
// final boolean[] same = new boolean[1000];
// for (int i = 0; i < 1000; i++) {
// final FacePair p =
// fold.trainingPairs.get(i);
// fInit[i] = p.loadFirst().asDoubleVector();
// sInit[i] = p.loadSecond().asDoubleVector();
// same[i] = p.same;
//
// for (int j = 0; j < fInit[i].length; j++) {
// if (Double.isInfinite(fInit[i][j]) || Double.isNaN(fInit[i][j]))
// throw new RuntimeException("" + fold.trainingPairs.get(i).firstFV);
// if (Double.isInfinite(sInit[i][j]) || Double.isNaN(sInit[i][j]))
// throw new RuntimeException("" + fold.trainingPairs.get(i).secondFV);
// }
// }
//
// System.out.println("LMDR Init");
// lmdr.recomputeBias(fInit, sInit, same);
// // lmdr.initialise(fInit, sInit, same);
// IOUtils.writeToFile(lmdr, new
// File("/Users/jon/Data/lfw/lmdr-matlabfvs-pcaw-init.bin"));
// // final LargeMarginDimensionalityReduction lmdr = IOUtils
// // .readFromFile(new File("/Users/jon/Data/lfw/lmdr-init.bin"));
//
// for (int i = 0; i < 1e6; i++) {
// if (i % 100 == 0)
// System.out.println("Iter " + i);
// final FacePair p = fold.trainingPairs.get(i);
// lmdr.step(p.loadFirst().asDoubleVector(),
// p.loadSecond().asDoubleVector(), p.same);
// }
// IOUtils.writeToFile(lmdr, new
// File("/Users/jon/Data/lfw/lmdr-matlabfvs-pcaw.bin"));
final LargeMarginDimensionalityReduction lmdr =
IOUtils.readFromFile(new
File("/Users/jon/Data/lfw/lmdr-matlabfvs-pcaw.bin"));
// final LargeMarginDimensionalityReduction lmdr = loadMatlabLMDR();
// final LargeMarginDimensionalityReduction lmdr = loadMatlabPCAW();
final double[][] first = new double[fold.testPairs.size()][];
final double[][] second = new double[fold.testPairs.size()][];
final boolean[] same = new boolean[fold.testPairs.size()];
for (int j = 0; j < same.length; j++) {
final FacePair p = fold.testPairs.get(j);
first[j] = p.loadFirst().asDoubleVector();
second[j] = p.loadSecond().asDoubleVector();
same[j] = p.same;
}
// System.out.println("Current bias: " + lmdr.getBias());
// lmdr.recomputeBias(first, second, same);
// System.out.println("Best bias: " + lmdr.getBias());
double correct = 0;
double count = 0;
for (int j = 0; j < same.length; j++) {
final boolean pred = lmdr.classify(first[j],
second[j]);
if (pred == same[j])
correct++;
count++;
}
System.out.println(lmdr.getBias() + " " + (correct / count));
}
// private static double[] reorder(double[] in) {
// final double[] out = new double[in.length];
// final int D = 64;
// final int K = 512;
// for (int k = 0; k < K; k++) {
// for (int j = 0; j < D; j++) {
// out[k * D + j] = in[k * 2 * D + j];
// out[k * D + j + D * K] = in[k * 2 * D + j + D];
// }
// }
// return out;
// }
private static LargeMarginDimensionalityReduction loadMatlabLMDR() throws IOException {
final LargeMarginDimensionalityReduction lmdr = new LargeMarginDimensionalityReduction(128);
final MatFileReader reader = new MatFileReader(new File("/Users/jon/lmdr.mat"));
final MLSingle W = (MLSingle) reader.getContent().get("W");
final MLSingle b = (MLSingle) reader.getContent().get("b");
lmdr.setBias(b.get(0, 0));
final Matrix proj = new Matrix(W.getM(), W.getN());
for (int j = 0; j < W.getN(); j++) {
for (int i = 0; i < W.getM(); i++) {
proj.set(i, j, W.get(i, j));
}
}
lmdr.setTransform(proj);
return lmdr;
}
private static LargeMarginDimensionalityReduction loadMatlabPCAW() throws IOException {
final LargeMarginDimensionalityReduction lmdr = new LargeMarginDimensionalityReduction(128);
final MatFileReader reader = new MatFileReader(new File("/Users/jon/pcaw.mat"));
final MLSingle W = (MLSingle) reader.getContent().get("proj");
lmdr.setBias(169.6264190673828);
final Matrix proj = new Matrix(W.getM(), W.getN());
for (int j = 0; j < W.getN(); j++) {
for (int i = 0; i < W.getM(); i++) {
proj.set(i, j, W.get(i, j));
}
}
lmdr.setTransform(proj);
return lmdr;
}
}
© 2015 - 2025 Weber Informatics LLC | Privacy Policy