com.sankalp.characterreader.neuralnetwork.TrainingData.java Source code

Introduction

Here is the source code for com.sankalp.characterreader.neuralnetwork.TrainingData.java

Source

/*
 * To change this license header, choose License Headers in Project Properties.
 * To change this template file, choose Tools | Templates
 * and open the template in the editor.
 */

package com.sankalp.characterreader.neuralnetwork;

import com.sankalp.characterreader.neuralnetwork.TrainingData.Sample;
import javax.imageio.ImageIO;

import java.awt.image.BufferedImage;
import java.io.File;
import java.io.IOException;
import java.util.ArrayList;
import java.util.Collections;
import java.util.Iterator;
import java.util.List;

import org.apache.commons.io.FileUtils;

/**
 *
 * @author sankalpkulshrestha
 */
public class TrainingData implements Iterator<Sample> {

    private final int THRESHOLD = 100000;
    private File trainingDataFolder;
    private File outputFile;
    private int trainingSampleCount;
    private List<Sample> sampleList;
    private int currentCount = 0;
    private int currentIndex = 0;
    private int sampleIndex = 0;

    public TrainingData(File trainingDataFolder, int trainingSampleCount, File outputFile) throws IOException {
        this.trainingDataFolder = trainingDataFolder;
        this.outputFile = outputFile;
        this.trainingSampleCount = trainingSampleCount;
        sampleList = new ArrayList(trainingSampleCount < THRESHOLD ? trainingSampleCount : THRESHOLD);
        load();
    }

    private void load() throws IOException {
        int index;
        List<String> lineList = FileUtils.readLines(outputFile, "UTF-8");
        for (index = currentCount; index < (trainingSampleCount - currentCount < THRESHOLD
                ? trainingSampleCount - currentCount
                : THRESHOLD) + currentCount; index++) {
            sampleList.add(index - currentCount,
                    new Sample(new File(trainingDataFolder + "/" + Integer.toString(index) + ".jpg"),
                            Integer.parseInt(lineList.get(index).split(",")[1])));
        }
        sampleIndex = 0;
        currentCount = index;
    }

    public static class Sample {

        Integer[][] matrix;
        Integer[] expectedOutputArray;
        int expectedOutput;

        public Sample(BufferedImage image, int expectedOutputIndex) {
            matrix = new Integer[image.getHeight()][image.getWidth()];
            this.expectedOutput = expectedOutputIndex;
            this.expectedOutputArray = new Integer[] { 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 };
            this.expectedOutputArray[expectedOutputIndex] = 1;
            int col, row;
            for (col = 0; col < image.getWidth(); col++) {
                for (row = 0; row < image.getHeight(); row++) {
                    matrix[row][col] = image.getRGB(col, row);
                    if (matrix[row][col] == -1) {
                        matrix[row][col] = 1;
                    } else {
                        matrix[row][col] = 0;
                    }
                }
            }
        }

        public Sample(File imageFile, int expectedOutputIndex) {
            try {
                final BufferedImage image = ImageIO.read(imageFile);
                matrix = new Integer[image.getHeight()][image.getWidth()];
                this.expectedOutput = expectedOutputIndex;
                this.expectedOutputArray = new Integer[] { 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 };
                this.expectedOutputArray[expectedOutputIndex] = 1;
                int col, row;
                for (col = 0; col < image.getWidth(); col++) {
                    for (row = 0; row < image.getHeight(); row++) {
                        matrix[row][col] = image.getRGB(col, row);
                        if (matrix[row][col] == -1) {
                            matrix[row][col] = 1;
                        } else {
                            matrix[row][col] = 0;
                        }
                    }
                }
            } catch (IOException ex) {
                System.out.println(ex);
            }
        }
    }

    @Override
    public boolean hasNext() {
        return currentIndex < trainingSampleCount;
    }

    @Override
    public Sample next() {
        if (hasNext()) {
            if (sampleIndex < THRESHOLD) {
                sampleIndex++;
                currentIndex++;
                return sampleList.get(sampleIndex - 1);
            } else {
                try {
                    load();
                } catch (IOException ex) {
                    System.out.println(ex);
                }
                sampleIndex++;
                currentIndex++;
                return sampleList.get(sampleIndex - 1);
            }
        } else {
            return null;
        }
    }

    public void shuffle() {
        Collections.shuffle(sampleList);
    }

    public void reset() {
        currentCount = 0;
        currentIndex = 0;
        sampleIndex = 0;
    }

    @Override
    public void remove() {
        throw new UnsupportedOperationException("Not supported yet.");
        //To change body of generated methods, choose Tools | Templates.
    }

}