org.openimaj.demos.FVFWCheckPCAGMM Maven / Gradle / Ivy

Go to download
/**
 * 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.IOException;
import java.util.Arrays;

import org.openimaj.feature.FloatFV;
import org.openimaj.feature.local.list.MemoryLocalFeatureList;
import org.openimaj.image.feature.dense.gradient.dsift.FloatDSIFTKeypoint;
import org.openimaj.image.feature.local.aggregate.FisherVector;
import org.openimaj.math.matrix.algorithm.pca.PrincipalComponentAnalysis;
import org.openimaj.math.matrix.algorithm.pca.ThinSvdPrincipalComponentAnalysis;
import org.openimaj.math.statistics.distribution.DiagonalMultivariateGaussian;
import org.openimaj.math.statistics.distribution.MixtureOfGaussians;
import org.openimaj.math.statistics.distribution.MultivariateGaussian;
import org.openimaj.util.array.ArrayUtils;

import Jama.Matrix;

import com.jmatio.io.MatFileReader;
import com.jmatio.io.MatFileWriter;
import com.jmatio.types.MLArray;
import com.jmatio.types.MLDouble;
import com.jmatio.types.MLSingle;
import com.jmatio.types.MLStructure;

/**
 * 
 * @author Sina Samangooei ([email protected])
 */
public class FVFWCheckPCAGMM {

	private static final String GMM_MATLAB_FILE = "/Users/ss/Experiments/FVFW/data/gmm_512.mat";
	private static final String PCA_MATLAB_FILE = "/Users/ss/Experiments/FVFW/data/PCA_64.mat";
	private static final String[] FACE_DSIFTS = new String[] {
			"/Users/ss/Experiments/FVFW/data/Aaron_Eckhart_0001-pdfsift.bin"
	};

	public static void main(String[] args) throws IOException {
		final MixtureOfGaussians mog = loadMoG(new File(GMM_MATLAB_FILE));
		final PrincipalComponentAnalysis pca = loadPCA(new File(PCA_MATLAB_FILE));
		final FisherVector fisher = new FisherVector(mog, true, true);
		for (final String faceFile : FACE_DSIFTS) {
			final MemoryLocalFeatureList loadDSIFT = MemoryLocalFeatureList.read(new File(faceFile),
					FloatDSIFTKeypoint.class);
			projectPCA(loadDSIFT, pca);

			final FloatFV fvec = fisher.aggregate(loadDSIFT);
			System.out.println(String.format("%s: %s", faceFile, fvec));
			System.out.println("Writing...");

			final File out = new File(faceFile + ".fisher.mat");
			final MLArray data = toMLArray(fvec);
			new MatFileWriter(out, Arrays.asList(data));
		}
	}

	private static MLArray toMLArray(FloatFV fvec) {
		final MLDouble data = new MLDouble("fisherface", new int[] { fvec.values.length, 1 });
		for (int i = 0; i < fvec.values.length; i++) {
			data.set((double) fvec.values[i], i, 0);
		}
		return data;
	}

	private static void projectPCA(
			MemoryLocalFeatureList loadDSIFT,
			PrincipalComponentAnalysis pca)
	{
		for (final FloatDSIFTKeypoint kp : loadDSIFT) {
			kp.descriptor = ArrayUtils.convertToFloat(pca.project(ArrayUtils.convertToDouble(kp.descriptor)));
			final int nf = kp.descriptor.length;
			kp.descriptor = Arrays.copyOf(kp.descriptor, nf + 2);
			kp.descriptor[nf] = (kp.x / 125f) - 0.5f;
			kp.descriptor[nf + 1] = (kp.y / 160f) - 0.5f;
		}
		loadDSIFT.resetVecLength();
	}

	static class LoadedPCA extends ThinSvdPrincipalComponentAnalysis {

		public LoadedPCA(Matrix basis, double[] mean) {
			super(basis.getRowDimension());
			this.basis = basis;
			this.mean = mean;
		}

	}

	public static PrincipalComponentAnalysis loadPCA(File f) throws IOException {
		final MatFileReader reader = new MatFileReader(f);
		final MLSingle mean = (MLSingle) reader.getContent().get("mu");
		final MLSingle eigvec = (MLSingle) reader.getContent().get("proj");
		final Matrix basis = new Matrix(eigvec.getM(), eigvec.getN());
		final double[] meand = new double[eigvec.getN()];
		for (int j = 0; j < eigvec.getN(); j++) {
			// meand[i] = mean.get(i,0); ignore the means
			meand[j] = 0;
			for (int i = 0; i < eigvec.getM(); i++) {
				basis.set(i, j, eigvec.get(i, j));
			}
		}
		final PrincipalComponentAnalysis ret = new LoadedPCA(basis.transpose(), meand);
		return ret;
	}

	public static MixtureOfGaussians loadMoG(File f) throws IOException {
		final MatFileReader reader = new MatFileReader(f);
		final MLStructure codebook = (MLStructure) reader.getContent().get("codebook");

		final MLSingle mean = (MLSingle) codebook.getField("mean");
		final MLSingle variance = (MLSingle) codebook.getField("variance");
		final MLSingle coef = (MLSingle) codebook.getField("coef");

		final int n_gaussians = mean.getN();
		final int n_dims = mean.getM();

		final MultivariateGaussian[] ret = new MultivariateGaussian[n_gaussians];
		final double[] weights = new double[n_gaussians];
		for (int i = 0; i < n_gaussians; i++) {
			weights[i] = coef.get(i, 0);
			final DiagonalMultivariateGaussian d = new DiagonalMultivariateGaussian(n_dims);
			for (int j = 0; j < n_dims; j++) {
				d.mean.set(0, j, mean.get(j, i));
				d.variance[j] = variance.get(j, i);
			}
			ret[i] = d;
		}

		return new MixtureOfGaussians(ret, weights);
	}

}