Skip to content
Snippets Groups Projects
Commit 9e0cc8c9 authored by ik1g19's avatar ik1g19
Browse files

comment code

parent aa2a4985
No related branches found
No related tags found
No related merge requests found
Hide whitespace changes
Inline Side-by-side
visioncw3/src/main/java/uk/ac/soton/ecs/ik1g19/App.java
+ 117
− 3
View file @ 9e0cc8c9
......@@ -17,11 +17,18 @@ import org.openimaj.feature.FloatFV;
import org.openimaj.feature.SparseIntFV;
import org.openimaj.feature.local.LocalFeatureImpl;
import org.openimaj.feature.local.Location;
import org.openimaj.feature.local.SpatialLocation;
import org.openimaj.feature.local.data.LocalFeatureListDataSource;
import org.openimaj.feature.local.list.LocalFeatureList;
import org.openimaj.feature.local.list.MemoryLocalFeatureList;
import org.openimaj.image.FImage;
import org.openimaj.image.Image;
import org.openimaj.image.ImageUtilities;
import org.openimaj.image.analyser.ImageAnalyser;
import org.openimaj.image.feature.local.aggregate.BagOfVisualWords;
import org.openimaj.image.pixel.sampling.RectangleSampler;
import org.openimaj.image.processor.SinglebandImageProcessor;
import org.openimaj.math.geometry.shape.Rectangle;
import org.openimaj.ml.annotation.linear.LiblinearAnnotator;
import org.openimaj.ml.clustering.FloatCentroidsResult;
import org.openimaj.ml.clustering.assignment.HardAssigner;
......@@ -29,6 +36,7 @@ import org.openimaj.ml.clustering.kmeans.FloatKMeans;
import org.openimaj.util.pair.IntFloatPair;
import java.util.ArrayList;
import java.util.Iterator;
import java.util.List;
import java.util.Map;
......@@ -45,7 +53,7 @@ public class App {
VFSListDataset<FImage> testingData;
try {
trainingData = new VFSGroupDataset<FImage>(
"zip:C:/Users/isaac/Documents/GitHub Repos/visioncw3/visioncw3/data/training.zip", ImageUtilities.FIMAGE_READER);
"zip:http://comp3204.ecs.soton.ac.uk/cw/training.zip", ImageUtilities.FIMAGE_READER);
testingData = new VFSListDataset<FImage>(
"zip:C:/Users/isaac/Documents/GitHub Repos/visioncw3/visioncw3/data/testing.zip", ImageUtilities.FIMAGE_READER);
} catch (Exception e) {
......@@ -62,7 +70,7 @@ public class App {
//analyse patches of 8x8 sizes
PatchFeature patch = new PatchFeature(8);
PatchFeature patch = new PatchFeature(8f,4f, 4f);
//train quantiser on a sample of the data set
......@@ -74,11 +82,13 @@ public class App {
FeatureExtractor<SparseIntFV, FImage> extractor = new PatchExtractor(patch, assigner);
//train classification
LiblinearAnnotator<FImage, String> ann = new LiblinearAnnotator<FImage, String>(
extractor, LiblinearAnnotator.Mode.MULTICLASS, SolverType.L2R_L2LOSS_SVC, 1.0, 0.00001);
ann.train(splits.getTrainingDataset());
//evaluation
ClassificationEvaluator<CMResult<String>, String, FImage> eval =
new ClassificationEvaluator<CMResult<String>, String, FImage>(
ann, splits.getTestDataset(), new CMAnalyser<FImage, String>(CMAnalyser.Strategy.SINGLE));
......@@ -90,39 +100,75 @@ public class App {
}
/**
* @desc trains the quantiser using a sample of data and densely sampled patches
* @param sample sample of image data
* @param patch patch sampler
* @return trained assigner
*/
static HardAssigner<float[], float[], IntFloatPair> trainQuantiser(
Dataset<FImage> sample, PatchFeature<FImage> patch)
{
List<LocalFeatureList<LocalFeatureImpl<Location, FloatFV>>> allkeys = new ArrayList<>();
for (FImage rec : sample) {
FImage img = rec;
//densely sample patches and store
patch.analyseImage(img);
allkeys.add(patch.getPatchKeypoints());
}
//take first 1000 samples
if (allkeys.size() > 10000)
allkeys = allkeys.subList(0, 10000);
//k-means clustering
FloatKMeans km = FloatKMeans.createExact(500);
DataSource<float[]> datasource = new LocalFeatureListDataSource<LocalFeatureImpl<Location, FloatFV>, float[]>(allkeys);
DataSource<float[]> datasource =
new LocalFeatureListDataSource<LocalFeatureImpl<Location, FloatFV>, float[]>(allkeys);
FloatCentroidsResult result = km.cluster(datasource);
return result.defaultHardAssigner();
}
}
/**
* @desc feature extractor to sample patches as visual words
*/
class PatchExtractor implements FeatureExtractor<SparseIntFV, FImage> {
PatchFeature<FImage> patch;
HardAssigner<float[], float[], IntFloatPair> assigner;
/**
* @desc construct an extractor using a patch sampler and assigner
* @param patch a dense patch sampler
* @param assigner the trained quantiser
*/
public PatchExtractor(PatchFeature<FImage> patch, HardAssigner<float[], float[], IntFloatPair> assigner)
{
this.patch = patch;
this.assigner = assigner;
}
/**
* @desc extract visual words from image
* @param object FImage to extract from
* @return extracted visual words as a SpareIntFV
*/
public SparseIntFV extractFeature(FImage object) {
FImage image = object.getImage();
patch.analyseImage(image);
......@@ -131,4 +177,72 @@ class PatchExtractor implements FeatureExtractor<SparseIntFV, FImage> {
return bovw.aggregate(patch.getPatchKeypoints());
}
}
/**
* @desc image analyser for densely sampling patches across an image
*/
class PatchFeature <IMAGE extends Image<Float, IMAGE> & SinglebandImageProcessor.Processable<Float, FImage, IMAGE>> implements ImageAnalyser<IMAGE> {
float size;
float stepx;
float stepy;
private LocalFeatureList<LocalFeatureImpl<Location, FloatFV>> patchList;
/**
* @desc construct a patch sampler with given size and step
* @param size the size of the patch (a x a)
* @param stepx size of horizontal step
* @param stepy size of vertical step
*/
public PatchFeature(float size, float stepx, float stepy) {
this.size = size;
this.stepx = stepx;
this.stepy = stepy;
patchList = new MemoryLocalFeatureList<>();
}
public void analyseImage(IMAGE image) {
//clear patches of previous analysis
patchList.clear();
//sample patches using sliding window provided by RectangleSampler
RectangleSampler sampler = new RectangleSampler(image, stepx, stepy, size, size);
Iterator<Rectangle> rects = sampler.allRectangles().iterator();
while (rects.hasNext()) {
Rectangle rect = rects.next();
IMAGE rImg = image.extractROI(rect).normalise();
Float[] pvF = new Float[rImg.getWidth()* rImg.getHeight()];
rImg.getPixelVector(pvF);
float[] pvf = new float[rImg.getWidth()* rImg.getHeight()];
for (int i = 0; i < pvF.length; i++) { pvf[i] = pvF[i].floatValue(); }
FloatFV fv = new FloatFV(pvf);
patchList.add(new LocalFeatureImpl<Location, FloatFV>(new SpatialLocation(rect.x, rect.y),fv));
}
}
/**
* @desc returns densely sampled patches of analysed image
* @return densely sampled patches as LocalFeatureList
*/
public LocalFeatureList<LocalFeatureImpl<Location, FloatFV>> getPatchKeypoints() {
return patchList;
}
}
\ No newline at end of file
visioncw3/src/main/java/uk/ac/soton/ecs/ik1g19/PatchFeature.java
+ 51
− 37
View file @ 9e0cc8c9
......@@ -23,40 +23,54 @@ import org.openimaj.math.geometry.shape.Rectangle;
import org.openimaj.util.array.ArrayUtils;
import org.openimaj.util.pair.IntObjectPair;
public class PatchFeature <IMAGE extends Image<Float, IMAGE> & Processable<Float, FImage, IMAGE>> implements ImageAnalyser<IMAGE> {
float size;
private LocalFeatureList<LocalFeatureImpl<Location, FloatFV>> patchList;
public PatchFeature(float size) {
this.size = size;
patchList = new MemoryLocalFeatureList<>();
}
public void analyseImage(IMAGE image) {
patchList.clear();
RectangleSampler sampler = new RectangleSampler(image, 4f, 4f, size, size);
Iterator<Rectangle> rects = sampler.allRectangles().iterator();
while (rects.hasNext()) {
Rectangle rect = rects.next();
IMAGE rImg = image.extractROI(rect).normalise();
Float[] pvF = new Float[rImg.getWidth()* rImg.getHeight()];
rImg.getPixelVector(pvF);
float[] pvf = new float[rImg.getWidth()* rImg.getHeight()];
for (int i = 0; i < pvF.length; i++) { pvf[i] = pvF[i].floatValue(); }
FloatFV fv = new FloatFV(pvf);
patchList.add(new LocalFeatureImpl<Location, FloatFV>(new SpatialLocation(rect.x, rect.y),fv));
}
}
public LocalFeatureList<LocalFeatureImpl<Location, FloatFV>> getPatchKeypoints() {
return patchList;
}
}
///**
// * @desc image analyser for densely sampling patches across an image
// */
//public class PatchFeature <IMAGE extends Image<Float, IMAGE> & Processable<Float, FImage, IMAGE>> implements ImageAnalyser<IMAGE> {
// float size;
// float stepx;
// float stepy;
// private LocalFeatureList<LocalFeatureImpl<Location, FloatFV>> patchList;
//
// /**
// *
// * @param size
// * @param stepx
// * @param stepy
// */
// public PatchFeature(float size, float stepx, float stepy) {
// this.size = size;
// this.stepx = stepx;
// this.stepy = stepy;
// patchList = new MemoryLocalFeatureList<>();
// }
//
//
// public void analyseImage(IMAGE image) {
// patchList.clear();
//
// RectangleSampler sampler = new RectangleSampler(image, stepx, stepy, size, size);
//
// Iterator<Rectangle> rects = sampler.allRectangles().iterator();
//
// while (rects.hasNext()) {
// Rectangle rect = rects.next();
//
// IMAGE rImg = image.extractROI(rect).normalise();
//
// Float[] pvF = new Float[rImg.getWidth()* rImg.getHeight()];
// rImg.getPixelVector(pvF);
//
// float[] pvf = new float[rImg.getWidth()* rImg.getHeight()];
// for (int i = 0; i < pvF.length; i++) { pvf[i] = pvF[i].floatValue(); }
//
// FloatFV fv = new FloatFV(pvf);
//
// patchList.add(new LocalFeatureImpl<Location, FloatFV>(new SpatialLocation(rect.x, rect.y),fv));
// }
// }
//
// public LocalFeatureList<LocalFeatureImpl<Location, FloatFV>> getPatchKeypoints() {
// return patchList;
// }
//}
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Please register or to comment