boofcv.factory.feature.detect.interest.FactoryInterestPoint Maven / Gradle / Ivy

Go to download
Show more of this group Show more artifacts with this name
Show all versions of boofcv-feature Show documentation
BoofCV is an open source Java library for real-time computer vision and robotics applications.
There is a newer version: 1.1.7
/*
 * Copyright (c) 2011-2020, 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.factory.feature.detect.interest;

import boofcv.abst.feature.describe.ConfigSiftScaleSpace;
import boofcv.abst.feature.detect.extract.NonMaxLimiter;
import boofcv.abst.feature.detect.extract.NonMaxSuppression;
import boofcv.abst.feature.detect.interest.*;
import boofcv.abst.filter.derivative.ImageGradient;
import boofcv.abst.filter.derivative.ImageHessian;
import boofcv.alg.feature.detect.interest.*;
import boofcv.alg.filter.derivative.GImageDerivativeOps;
import boofcv.factory.feature.detect.extract.FactoryFeatureExtractor;
import boofcv.factory.filter.derivative.FactoryDerivative;
import boofcv.factory.transform.pyramid.FactoryPyramid;
import boofcv.struct.image.ImageGray;
import boofcv.struct.pyramid.PyramidFloat;

import javax.annotation.Nullable;

/**
 * Factory for creating interest point detectors which conform to the {@link InterestPointDetector}
 * interface 
 * 
 * NOTE: Higher level interface than {@link GeneralFeatureDetector}.  This will automatically
 * compute image derivatives across scale space as needed, unlike GeneralFeatureDetector which
 * just detects features at a particular scale and requires image derivatives be passed in.
 * 
 *
 * @author Peter Abeles
 * @see FactoryFeatureExtractor
 */
public class FactoryInterestPoint {

	public static , D extends ImageGray>
	InterestPointDetector generic( ConfigDetectInterestPoint config , Class inputType, @Nullable Class derivType )
	{
		switch(config.type) {
			case FAST_HESSIAN: return FactoryInterestPoint.fastHessian(config.fastHessian,inputType);
			case SIFT: return FactoryInterestPoint.sift(config.scaleSpaceSift,config.sift,inputType);
			case POINT: {
				if( derivType == null )
					derivType = GImageDerivativeOps.getDerivativeType(inputType);
				GeneralFeatureDetector alg = FactoryDetectPoint.create(config.point,inputType,derivType);
				return FactoryInterestPoint.wrapPoint(alg, config.point.scaleRadius,inputType, derivType);
			}
			default:
				throw new IllegalArgumentException("Unknown detector");
		}
	}

	/**
	 * Wraps {@link GeneralFeatureDetector} inside an {@link InterestPointDetector}.
	 *
	 * @param feature Feature detector.
	 * @param scaleRadius Radius of descriptor used by scale invariant features
	 * @param inputType Image type of input image.
	 * @param derivType Image type for gradient.
	 * @return The interest point detector.
	 */
	public static , D extends ImageGray>
	InterestPointDetector wrapPoint(GeneralFeatureDetector feature, double scaleRadius , Class inputType, Class derivType) {
		if( derivType == null )
			derivType = GImageDerivativeOps.getDerivativeType(inputType);

		ImageGradient gradient = null;
		ImageHessian hessian = null;

		if (feature.getRequiresGradient() || feature.getRequiresHessian())
			gradient = FactoryDerivative.sobel(inputType, derivType);
		if (feature.getRequiresHessian())
			hessian = FactoryDerivative.hessianSobel(derivType);

		return new GeneralToInterestPoint<>(feature, gradient, hessian, scaleRadius, derivType);
	}

	/**
	 * Wraps {@link FeatureLaplacePyramid} inside an {@link InterestPointDetector}.
	 *
	 * @param feature   Feature detector.
	 * @param scales    Scales at which features are detected at.
	 * @param pyramid   Should it be constructed as a pyramid or scale-space
	 * @param inputType Image type of input image.
	 * @return The interest point detector.
	 */
	public static , D extends ImageGray>
	InterestPointDetector wrapDetector(FeatureLaplacePyramid feature,
										  double[] scales, boolean pyramid,
										  Class inputType) {

		PyramidFloat ss;

		if( pyramid )
			ss = FactoryPyramid.scaleSpacePyramid(scales, inputType);
		else
			ss = FactoryPyramid.scaleSpace(scales, inputType);

		return new WrapFLPtoInterestPoint<>(feature, ss);
	}

	/**
	 * Wraps {@link FeaturePyramid} inside an {@link InterestPointDetector}.
	 *
	 * @param feature   Feature detector.
	 * @param scales    Scales at which features are detected at.
	 * @param pyramid   Should it be constructed as a pyramid or scale-space
	 * @param inputType Image type of input image.
	 * @return The interest point detector.
	 */
	public static , D extends ImageGray>
	InterestPointDetector wrapDetector(FeaturePyramid feature,
										  double[] scales, boolean pyramid,
										  Class inputType) {

		PyramidFloat ss;

		if( pyramid )
			ss = FactoryPyramid.scaleSpacePyramid(scales, inputType);
		else
			ss = FactoryPyramid.scaleSpace(scales, inputType);

		return new WrapFPtoInterestPoint<>(feature, ss);
	}

	/**
	 * Creates a {@link FastHessianFeatureDetector} detector which is wrapped inside
	 * an {@link InterestPointDetector}
	 *
	 * @param config Configuration for detector.  Pass in null for default options.
	 * @return The interest point detector.
	 * @see FastHessianFeatureDetector
	 */
	public static >
	InterestPointDetector fastHessian( ConfigFastHessian config, Class imageType ) {
		return new WrapFHtoInterestPoint(FactoryInterestPointAlgs.fastHessian(config),imageType);
	}

	public static >
	InterestPointDetector sift(ConfigSiftScaleSpace configSS ,
								  ConfigSiftDetector configDet , Class imageType ) {

		if( configSS == null )
			configSS = new ConfigSiftScaleSpace();
		if( configDet == null )
			configDet = new ConfigSiftDetector();

		SiftScaleSpace scaleSpace =
				new SiftScaleSpace(configSS.firstOctave,configSS.lastOctave,configSS.numScales,configSS.sigma0);
		NonMaxSuppression nonmax = FactoryFeatureExtractor.nonmax(configDet.extract);
		NonMaxLimiter limiter = new NonMaxLimiter(nonmax,configDet.maxFeaturesPerScale);
		SiftDetector detector = new SiftDetector(scaleSpace,configDet.edgeR,limiter);

		return new WrapSiftDetector<>(detector, imageType);
	}
}