1 /**
2 * Copyright (c) 2011, The University of Southampton and the individual contributors.
3 * All rights reserved.
4 *
5 * Redistribution and use in source and binary forms, with or without modification,
6 * are permitted provided that the following conditions are met:
7 *
8 * * Redistributions of source code must retain the above copyright notice,
9 * this list of conditions and the following disclaimer.
10 *
11 * * Redistributions in binary form must reproduce the above copyright notice,
12 * this list of conditions and the following disclaimer in the documentation
13 * and/or other materials provided with the distribution.
14 *
15 * * Neither the name of the University of Southampton nor the names of its
16 * contributors may be used to endorse or promote products derived from this
17 * software without specific prior written permission.
18 *
19 * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND
20 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
21 * WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
22 * DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR
23 * ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
24 * (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
25 * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON
26 * ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
27 * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
28 * SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
29 */
30 package org.openimaj.experiment.gmm.retrieval;
31
32 import org.openimaj.feature.FeatureVector;
33 import org.openimaj.feature.local.LocalFeature;
34 import org.openimaj.feature.local.list.LocalFeatureList;
35 import org.openimaj.math.statistics.distribution.MixtureOfGaussians;
36 import org.openimaj.ml.gmm.GaussianMixtureModelEM;
37 import org.openimaj.ml.gmm.GaussianMixtureModelEM.CovarianceType;
38 import org.openimaj.util.array.ArrayUtils;
39 import org.openimaj.util.function.Function;
40
41 import Jama.Matrix;
42
43 /**
44 * This function turns a list of features to a gaussian mixture model
45 *
46 * @author Sina Samangooei (ss@ecs.soton.ac.uk)
47 */
48 public class GMMFromFeatures implements Function<
49 LocalFeatureList<? extends LocalFeature<?,? extends FeatureVector>>,
50 MixtureOfGaussians
51 >
52 {
53
54 /**
55 * default number of guassians to train agains
56 */
57 public static final int DEFAULT_COMPONENTS = 10;
58 /**
59 * default covariance type
60 */
61 public static final CovarianceType DEFAULT_COVARIANCE = GaussianMixtureModelEM.CovarianceType.Spherical;
62
63 private GaussianMixtureModelEM gmm;
64 /**
65 * Defaults to {@link #DEFAULT_COMPONENTS} and
66 */
67 public GMMFromFeatures() {
68 this.gmm = new GaussianMixtureModelEM(DEFAULT_COMPONENTS, DEFAULT_COVARIANCE);
69 }
70
71 /**
72 * @param type
73 */
74 public GMMFromFeatures(CovarianceType type) {
75 this.gmm = new GaussianMixtureModelEM(DEFAULT_COMPONENTS, type);
76 }
77
78 /**
79 * @param nComps
80 */
81 public GMMFromFeatures(int nComps) {
82 this.gmm = new GaussianMixtureModelEM(nComps, DEFAULT_COVARIANCE);
83 }
84
85 /**
86 * @param nComps
87 * @param type
88 */
89 public GMMFromFeatures(int nComps,CovarianceType type) {
90 this.gmm = new GaussianMixtureModelEM(nComps, type);
91 }
92
93 @Override
94 public MixtureOfGaussians apply(LocalFeatureList<? extends LocalFeature<?,? extends FeatureVector>> features) {
95 System.out.println("Creating double array...");
96 double[][] doubleFeatures = new double[features.size()][];
97 int i = 0;
98 for (LocalFeature<?,?> localFeature : features) {
99 doubleFeatures[i] = ArrayUtils.divide(localFeature.getFeatureVector().asDoubleVector(), 128);
100 i++;
101 }
102 System.out.println(String.format("Launching EM with double array: %d x %d",doubleFeatures.length,doubleFeatures[0].length));
103 return this.gmm.estimate(new Matrix(doubleFeatures));
104 }
105
106
107
108 }