boofcv.alg.scene.nister2006.TupleMapDistanceNorm Maven / Gradle / Ivy

/*
 * Copyright (c) 2021, Peter Abeles. All Rights Reserved.
 *
 * This file is part of BoofCV (http://boofcv.org).
 *
 * Licensed under the Apache License, Version 2.0 (the "License");
 * you may not use this file except in compliance with the License.
 * You may obtain a copy of the License at
 *
 *   http://www.apache.org/licenses/LICENSE-2.0
 *
 * Unless required by applicable law or agreed to in writing, software
 * distributed under the License is distributed on an "AS IS" BASIS,
 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 * See the License for the specific language governing permissions and
 * limitations under the License.
 */

package boofcv.alg.scene.nister2006;

import boofcv.misc.BoofMiscOps;
import org.ddogleg.struct.DogArray_F32;

/**
 * Generalized way for normalizing and computing the distance between two sparse descriptors in a map
 * format. Intended for use with {@link RecognitionVocabularyTreeNister2006}. Uses efficient distance
 * formula from [1].
 *
 * 
 * [1] Nister, David, and Henrik Stewenius. "Scalable recognition with a vocabulary tree."
 * 2006 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR'06). Vol. 2. Ieee, 2006.
 * 
 *
 * @author Peter Abeles
 */
public interface TupleMapDistanceNorm {
	/**
	 * Normalizes the descriptor. Computes the norm then divides each element by the norm.
	 */
	void normalize( DogArray_F32 weights );

	/**
	 * Incremental update to the distance. Initially the distance is set to 2.0, then for every word that
	 * is present in both descriptors add this value to it.
	 *
	 * This only works for a subclass of normalizations. See [1] for details.
	 */
	float distanceUpdate( float valA, float valB );

	/** Create a new instance that is thread safe, i.e. read only settings can be shared */
	TupleMapDistanceNorm newInstanceThread();

	/** Distance functions that are supported */
	enum Types {L1, L2}

	/**
	 * L1-norm for scoring
	 */
	class L1 implements TupleMapDistanceNorm {
		@Override public void normalize( DogArray_F32 weights ) {
			float norm = 0;
			for (int i = 0; i < weights.size; i++) {
				norm += weights.get(i);
			}
			BoofMiscOps.checkTrue(norm != 0.0f, "Norm value is zero. Something went very wrong");

			for (int i = 0; i < weights.size; i++) {
				weights.data[i] /= norm;
			}
		}

		@Override public float distanceUpdate( float valA, float valB ) {
			return Math.abs(valA - valB) - valA - valB;
		}

		@Override public TupleMapDistanceNorm newInstanceThread() {
			return new L1();
		}
	}

	/**
	 * L2-norm for scoring
	 */
	class L2 implements TupleMapDistanceNorm {
		@Override public void normalize( DogArray_F32 weights ) {
			float norm = 0;
			for (int i = 0; i < weights.size; i++) {
				float value = weights.data[i];
				norm += value*value;
			}
			norm = (float)Math.sqrt(norm);
			BoofMiscOps.checkTrue(norm != 0.0, "Sum of weights is zero. Something went very wrong");

			for (int i = 0; i < weights.size; i++) {
				weights.data[i] /= norm;
			}
		}

		@Override public float distanceUpdate( float valA, float valB ) {
			return -2.0f*valA*valB;
		}

		@Override public TupleMapDistanceNorm newInstanceThread() {
			return new L2();
		}
	}
}