NyARHistogramAnalyzer_DiscriminantThreshold.java

/* 
 * PROJECT: NyARToolkit(Extension)
 * --------------------------------------------------------------------------------
 * The NyARToolkit is Java edition ARToolKit class library.
 * Copyright (C)2008-2009 Ryo Iizuka
 *
 * This program is free software: you can redistribute it and/or modify
 * it under the terms of the GNU General Public License as published by
 * the Free Software Foundation, either version 3 of the License, or
 * (at your option) any later version.
 * 
 * This program is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 * GNU General Public License for more details.
 *
 * You should have received a copy of the GNU General Public License
 * along with this program.  If not, see <http://www.gnu.org/licenses/>.
 * 
 * For further information please contact.
 *	http://nyatla.jp/nyatoolkit/
 *	<airmail(at)ebony.plala.or.jp> or <nyatla(at)nyatla.jp>
 * 
 */
package jp.nyatla.nyartoolkit.core.analyzer.histogram;

import jp.nyatla.nyartoolkit.core.types.NyARHistogram;

/**
 * このクラスは、判別法を用いて敷居値を求める機能を提供します。
 * <p>memo:画素数が2048^2に満たない場合は、fixedint(24-8)で計算できます。</p>
 */
public class NyARHistogramAnalyzer_DiscriminantThreshold implements INyARHistogramAnalyzer_Threshold
{
	private double _score;
	/**
	 * この関数は、判別法を用いて敷居値を1個求めます。敷居値の範囲は、i_histogram引数の範囲と同じです。
	 * 関数は、thisのプロパティを更新します。
	 */
	public int getThreshold(NyARHistogram i_histogram)
	{
		int[] hist=i_histogram.data;
		int n=i_histogram.length;
		int da,sa,db,sb,dt,pt,st;
		int i;		
		int th=0;
		//後で使う
		dt=pt=0;
		for(i=0;i<n;i++){
			int h=hist[i];
			dt+=h*i;
			pt+=h*i*i;//正規化の時に使う定数
		}
		st=i_histogram.total_of_data;
		//Low側(0<=i<=n-2)
		da=dt;
		sa=st;
		//High側(i=n-1)
		db=sb=0;		
		
		double max=-1;
		double max_mt=0;
		//各ヒストグラムの分離度を計算する(1<=i<=n-1の範囲で評価)
		for(i=n-1;i>0;i--){
			//次のヒストグラムを計算
			int hist_count=hist[i];
			int hist_val=hist_count*i;
			da-=hist_val;
			sa-=hist_count;
			db+=hist_val;
			sb+=hist_count;
			
			//クラス間分散を計算
			double dv=(sa+sb);
			double mt=(double)(da+db)/dv;
			double ma=(sa!=0?((double)da/(double)sa):0)-mt;
			double mb=(sb!=0?((double)db/(double)sb):0)-mt;
			double kai=((double)(sa*(ma*ma)+sb*(mb*mb)))/dv;
			if(max<kai){
				max_mt=mt;
				max=kai;
				th=i;
			}
			//System.out.println(kai);
		}
		//max_mtを元に正規化
		this._score=max/((double)(pt+max_mt*max_mt*st-2*max_mt*dt)/st);//129,0.8888888888888887
		return th;
	}
	/**
	 * 最後に実行した{@link #getThreshold}のスコアを返却します。
	 * @return
	 * スコア値。範囲は0<=n<=1.0の間です。
	 */
	public final double getLastScore()
	{
		return this._score;
	}
	/**
	 * デバック用関数
	 * @param args
	 * main関数引数
	 */
	public static void main(String[] args)
	{
		NyARHistogram data=new NyARHistogram(256);
		for(int i=0;i<256;i++){
			data.data[i]=128-i>0?128-i:i-128;
		}
		data.total_of_data=data.getTotal(0,255);
		NyARHistogramAnalyzer_DiscriminantThreshold an=new NyARHistogramAnalyzer_DiscriminantThreshold();
		int th=an.getThreshold(data);
		System.out.print(th);
		return;
	}
}