Gui.java

import java.awt.Color;
import java.awt.Container;
import java.awt.Dimension;
import java.awt.Graphics;
import java.awt.event.ActionEvent;
import java.awt.event.ActionListener;
import java.awt.event.MouseEvent;
import java.awt.event.MouseMotionListener;
import java.io.BufferedReader;
import java.io.FileInputStream;
import java.io.InputStreamReader;
import java.util.Vector;

import javax.swing.BoxLayout;
import javax.swing.JButton;
import javax.swing.JComboBox;
import javax.swing.JFrame;
import javax.swing.JPanel;
import javax.swing.JTextField;

/*
 * 作成日: 2004/01/18
 *　$Id: Gui.java,v 1.4 2004/01/19 09:40:57 matsu Exp $
 */

/**
 * @author matsu
 *
 * 文字認識GUIとControl
 */
public class Gui
  extends JFrame
  implements MouseMotionListener, ActionListener {

  /** 認識文字表示領域*/
  private JTextField text = new JTextField();

  /** 文字認識パネル */
  private JPanel field_panel;

  /** mode */
  private JButton learning_button = new JButton("Learning mode");

  /** mode */
  private JButton recognition_button = new JButton("Recognition mode");

  /** リセットボタン */
  private JButton reset_button = new JButton("RESET");

  /** OKボタン */
  private JButton ok_button = new JButton("OK");

  /** 答え表示パネル */
  private JPanel text_panel = new JPanel();

  /** 文字選択パネル */
  private JPanel select_panel;

  /** セレクトボックス */
  private JComboBox select;

  /** Container */
  private Container c;

  /** NeuralNetwork */
  private NeuralNetwork neural;

  /** 学習データ */
  Vector vector = new Vector();

  // アルファベット配列
  private static final String[] alpha_array =
    new String[] {
      "A",
      "B",
      "C",
      "D",
      "E",
      "F",
      "G",
      "H",
      "I",
      "J",
      "K",
      "L",
      "M",
      "N",
      "O",
      "P",
      "Q",
      "R",
      "S",
      "T",
      "U",
      "V",
      "W",
      "X",
      "Y",
      "Z" };

  /** 升目 */
  private static final int x = 8;
  /** 升目 */
  private static final int y = 8;

  /** 一升の大きさ */
  private static final int x_width = 20;

  /** 一升の大きさ */
  private static final int y_width = 20;

  /** 何個候補を表示するか */
  private static final int DISP = 5;

  /** 学習データ */
  private static final String learning_file = "learning.txt";

  /** 入力情報格納 */
  private int[][] input = new int[x][y];

  /** モード定義 */
  public static final int LEARNING = 0;
  public static final int RECOGNITION = 1;

  /** モード */
  private int mode = 0;

  /** ファイルから学習するか */
  private static boolean LEARN_FROM_FILE = false;

  /**
   * main
   * @param args
   */
  public static void main(String[] args) {
    //Create and set up the window.
    Gui frame = new Gui();
    frame.setDefaultCloseOperation(JFrame.EXIT_ON_CLOSE);

    //Display the window.
    frame.setVisible(true);
  }

  /**
   * GUI作成
   *
   */
  public Gui() {
    super("Pattern learning.");

    init();

    // ニューラルネットワーク
    neural = new NeuralNetwork();
    if(LEARN_FROM_FILE == true)
      study();

    // サイズ設定
    //setBounds(50, 50, 300, 500);
    //setSize(300,500);

    // Containerを取得
    c = getContentPane();
    c.setLayout(new BoxLayout(c, BoxLayout.Y_AXIS));

    c.add(makeButtonPane());
    c.add(makeSelectBox());
    c.add(makeField());
    c.add(makeButton());

    makeDisplayLine();

    //validate();
    pack();

  }

  /**
   * 初期化
   * @return
   */
  public void init() {
    for (int i = 0; i < input.length; i++) {
      for (int j = 0; j < input[i].length; j++)
        input[i][j] = 0;
    }
  }

  /**
   * モード選択
   * @return
   */
  public Container makeButtonPane() {
    JPanel panel = new JPanel();

    learning_button.addActionListener(this);
    recognition_button.addActionListener(this);

    panel.add(learning_button);
    panel.add(recognition_button);

    return panel;
  }

  /**
   * 学習させる文字選択
   * @return
   */
  public Container makeSelectBox() {
    select_panel = new JPanel();

    select = new JComboBox(alpha_array);

    select_panel.add(select);
    //panel.setPreferredSize(new Dimension(100, 50));
    //panel.setBackground(Color.blue);

    return select_panel;
  }

  /**
   * 結果表示領域
   * @return
   */
  public Container makeDisplayLine() {

    text.setPreferredSize(new Dimension(250, 20));

    text_panel.add(this.text);

    return text_panel;
  }

  /**
   * 入力フィールド
   * @return
   */
  public Container makeField() {
    field_panel = new JPanel() {
      public void paintComponent(Graphics g) {
        // 升目描画
        for (int i = 0; i <= x; i++) {
          // 縦線
          g.drawLine(i * x_width, 0, i * x_width, y * y_width);
          // 横線
          g.drawLine(0, i * x_width, x * x_width, i * x_width);
        }

        // 色描画
        for (int i = 0; i < x; i++) {
          for (int j = 0; j < y; j++) {
            if (input[i][j] == 0) {
              //白
              g.setColor(Color.WHITE);
            } else {
              g.setColor(Color.BLACK);

            }

            g.fillRect(
                j * y_width + 1,
                i * x_width + 1,
                x_width - 1,
                y_width - 1);

            g.setColor(Color.BLACK);
          }
        }

      }
    };
    field_panel.addMouseMotionListener(this);
    field_panel.setPreferredSize(new Dimension(x * x_width, y * y_width));
    return field_panel;
  }

  /**
   * OKボタン
   * @return
   */
  public Container makeButton() {
    JPanel panel = new JPanel();

    reset_button.addActionListener(this);
    ok_button.addActionListener(this);

    panel.add(reset_button);
    panel.add(ok_button);
    return panel;
  }

  /**
   * モードを変える
   *
   */
  public void changeMode(int next_mode) {

    c.remove(1);

    if (next_mode == LEARNING) {
      c.add(select_panel, 1);
    } else if (next_mode == RECOGNITION) {
      c.add(text_panel, 1);
    }

    pack();
    mode = next_mode;
  }

  /**
   * 学習
   */
  public void study() {
    Log.info("<-- Learn from data");

    Vector _vector = new Vector();

    try {
      // ファイルから読み込み
      BufferedReader br =
        new BufferedReader(
            new InputStreamReader(new FileInputStream(learning_file)));

      // 学習データ作成
      String msg;
      int i = 0;
      while ((msg = br.readLine()) != null) {
        for (int j = 0; j < msg.length(); j++) {
          double[] _teach = new double[alpha_array.length];
          double[] _input = new double[x * y];

          //init
          for (int h = 0; h < _teach.length; h++) {
            _teach[h] = 0;
            if (j == i)
              _teach[h] = 1;
          }
          for (int h = 0; h < x * y; h++) {
            _input[h] = Double.parseDouble(String.valueOf(msg.charAt(h)));
          }

          _vector.add(new Dataset(_input, _teach));

        }
        i++;
      }

    } catch (Exception e) {
      e.printStackTrace();
    }

    neural.learn(_vector);



    Log.info("-->");

  }

  /**
   * マウスアクション
   */
  public void mouseDragged(MouseEvent e) {

    // クリック場所識別
    if (e.getSource().equals(field_panel)) {
      if (e.getX() < x * x_width
          && e.getY() < y * y_width
          && 0 < e.getX()
          && 0 < e.getY()) {

        input[e.getY() / y_width][e.getX() / x_width] = 1;
        field_panel.repaint();
        //Log.debug(e.getX()/x_width+":"+e.getY()/y_width);
          }
    }
  }

  /**
   * マウスアクション
   */
  public void mouseMoved(MouseEvent e) {
    //Log.debug(e.toString());
  }

  /**
   * ボタンアクション
   */
  public void actionPerformed(ActionEvent e) {
    // Reset
    if (e.getSource().equals(reset_button)) {
      init();
      field_panel.repaint();
    }

    // OK
    else if (e.getSource().equals(ok_button)) {
      if (mode == LEARNING) {

        double[] _teach = new double[alpha_array.length];
        double[] _input = new double[x * y];

        // Inputデータ
        for (int i = 0; i < x; i++)
          for (int j = 0; j < y; j++) {
            _input[i * x + j] = input[i][j];
            System.out.print(input[i][j]);
          }

        // 教師データ
        for (int i = 0; i < alpha_array.length; i++) {
          _teach[i] = 0;
          if (select.getSelectedIndex() == i)
            _teach[i] = 1;
        }

        Dataset dataset = new Dataset(_input, _teach);

        vector.add(dataset);

        // 学習
        neural.learn(vector);

        init();
        field_panel.repaint();

      } else if (mode == RECOGNITION) {
        double[] _input = new double[x * y];

        // Inputデータ
        for (int i = 0; i < x; i++)
          for (int j = 0; j < y; j++)
            _input[i * x + j] = input[i][j];

        Dataset dataset = new Dataset(_input);

        // 認識
        neural.forward(dataset);

        // 結果取得
        double[] result = neural.getResult();

        // Sort
        int current = 0;
        int[] key = new int[result.length];

        // key init
        for (int i = 0; i < result.length; i++) {
          key[i] = i;
        }

        while (current < result.length) {
          for (int search = current + 1;
              search < result.length;
              search++) {
            if (result[current] < result[search]) {
              double tmp = result[current];
              result[current] = result[search];
              result[search] = tmp;

              int tmp_int = key[current];
              key[current] = key[search];
              key[search] = tmp_int;
            }
              }
          current++;
        }

        StringBuffer disp_text = new StringBuffer();
        for (int i = 0; i < DISP; i++) {
          disp_text.append(alpha_array[key[i]] + " ");
        }
        text.setText(disp_text.toString());
        pack();

        Log.info(alpha_array[key[0]]);
      }



    }
    // mode 
    else if (e.getSource().equals(learning_button)) {
      changeMode(LEARNING);
    }
    // mode
    else if (e.getSource().equals(recognition_button)) {
      changeMode(RECOGNITION);
    }
  }
}