1 Feature Detection and Description
2 =================================
8 * An example explaining keypoint detection and description can be found at opencv_source_code/samples/cpp/descriptor_extractor_matcher.cpp
12 Detects corners using the FAST algorithm
14 .. ocv:function:: void FAST( InputArray image, vector<KeyPoint>& keypoints, int threshold, bool nonmaxSuppression=true )
15 .. ocv:function:: void FAST( InputArray image, vector<KeyPoint>& keypoints, int threshold, bool nonmaxSuppression, int type )
17 .. ocv:pyfunction:: cv2.FastFeatureDetector([, threshold[, nonmaxSuppression]]) -> <FastFeatureDetector object>
18 .. ocv:pyfunction:: cv2.FastFeatureDetector(threshold, nonmaxSuppression, type) -> <FastFeatureDetector object>
19 .. ocv:pyfunction:: cv2.FastFeatureDetector.detect(image[, mask]) -> keypoints
22 :param image: grayscale image where keypoints (corners) are detected.
24 :param keypoints: keypoints detected on the image.
26 :param threshold: threshold on difference between intensity of the central pixel and pixels of a circle around this pixel.
28 :param nonmaxSuppression: if true, non-maximum suppression is applied to detected corners (keypoints).
30 :param type: one of the three neighborhoods as defined in the paper: ``FastFeatureDetector::TYPE_9_16``, ``FastFeatureDetector::TYPE_7_12``, ``FastFeatureDetector::TYPE_5_8``
32 Detects corners using the FAST algorithm by [Rosten06]_.
34 .. note:: In Python API, types are given as ``cv2.FAST_FEATURE_DETECTOR_TYPE_5_8``, ``cv2.FAST_FEATURE_DETECTOR_TYPE_7_12`` and ``cv2.FAST_FEATURE_DETECTOR_TYPE_9_16``. For corner detection, use ``cv2.FAST.detect()`` method.
37 .. [Rosten06] E. Rosten. Machine Learning for High-speed Corner Detection, 2006.
40 BriefDescriptorExtractor
41 ------------------------
42 .. ocv:class:: BriefDescriptorExtractor : public DescriptorExtractor
44 Class for computing BRIEF descriptors described in a paper of Calonder M., Lepetit V.,
45 Strecha C., Fua P. *BRIEF: Binary Robust Independent Elementary Features* ,
46 11th European Conference on Computer Vision (ECCV), Heraklion, Crete. LNCS Springer, September 2010. ::
48 class BriefDescriptorExtractor : public DescriptorExtractor
51 static const int PATCH_SIZE = 48;
52 static const int KERNEL_SIZE = 9;
54 // bytes is a length of descriptor in bytes. It can be equal 16, 32 or 64 bytes.
55 BriefDescriptorExtractor( int bytes = 32 );
57 virtual void read( const FileNode& );
58 virtual void write( FileStorage& ) const;
59 virtual int descriptorSize() const;
60 virtual int descriptorType() const;
61 virtual int defaultNorm() const;
68 * A complete BRIEF extractor sample can be found at opencv_source_code/samples/cpp/brief_match_test.cpp
73 .. ocv:class:: MSER : public FeatureDetector
75 Maximally stable extremal region extractor. ::
77 class MSER : public CvMSERParams
80 // default constructor
82 // constructor that initializes all the algorithm parameters
83 MSER( int _delta, int _min_area, int _max_area,
84 float _max_variation, float _min_diversity,
85 int _max_evolution, double _area_threshold,
86 double _min_margin, int _edge_blur_size );
87 // runs the extractor on the specified image; returns the MSERs,
88 // each encoded as a contour (vector<Point>, see findContours)
89 // the optional mask marks the area where MSERs are searched for
90 void operator()( const Mat& image, vector<vector<Point> >& msers, const Mat& mask ) const;
93 The class encapsulates all the parameters of the MSER extraction algorithm (see
94 http://en.wikipedia.org/wiki/Maximally_stable_extremal_regions). Also see http://code.opencv.org/projects/opencv/wiki/MSER for useful comments and parameters description.
98 * (Python) A complete example showing the use of the MSER detector can be found at opencv_source_code/samples/python2/mser.py
103 .. ocv:class:: ORB : public Feature2D
105 Class implementing the ORB (*oriented BRIEF*) keypoint detector and descriptor extractor, described in [RRKB11]_. The algorithm uses FAST in pyramids to detect stable keypoints, selects the strongest features using FAST or Harris response, finds their orientation using first-order moments and computes the descriptors using BRIEF (where the coordinates of random point pairs (or k-tuples) are rotated according to the measured orientation).
107 .. [RRKB11] Ethan Rublee, Vincent Rabaud, Kurt Konolige, Gary R. Bradski: ORB: An efficient alternative to SIFT or SURF. ICCV 2011: 2564-2571.
113 .. ocv:function:: ORB::ORB(int nfeatures = 500, float scaleFactor = 1.2f, int nlevels = 8, int edgeThreshold = 31, int firstLevel = 0, int WTA_K=2, int scoreType=ORB::HARRIS_SCORE, int patchSize=31)
115 .. ocv:pyfunction:: cv2.ORB([, nfeatures[, scaleFactor[, nlevels[, edgeThreshold[, firstLevel[, WTA_K[, scoreType[, patchSize]]]]]]]]) -> <ORB object>
118 :param nfeatures: The maximum number of features to retain.
120 :param scaleFactor: Pyramid decimation ratio, greater than 1. ``scaleFactor==2`` means the classical pyramid, where each next level has 4x less pixels than the previous, but such a big scale factor will degrade feature matching scores dramatically. On the other hand, too close to 1 scale factor will mean that to cover certain scale range you will need more pyramid levels and so the speed will suffer.
122 :param nlevels: The number of pyramid levels. The smallest level will have linear size equal to ``input_image_linear_size/pow(scaleFactor, nlevels)``.
124 :param edgeThreshold: This is size of the border where the features are not detected. It should roughly match the ``patchSize`` parameter.
126 :param firstLevel: It should be 0 in the current implementation.
128 :param WTA_K: The number of points that produce each element of the oriented BRIEF descriptor. The default value 2 means the BRIEF where we take a random point pair and compare their brightnesses, so we get 0/1 response. Other possible values are 3 and 4. For example, 3 means that we take 3 random points (of course, those point coordinates are random, but they are generated from the pre-defined seed, so each element of BRIEF descriptor is computed deterministically from the pixel rectangle), find point of maximum brightness and output index of the winner (0, 1 or 2). Such output will occupy 2 bits, and therefore it will need a special variant of Hamming distance, denoted as ``NORM_HAMMING2`` (2 bits per bin). When ``WTA_K=4``, we take 4 random points to compute each bin (that will also occupy 2 bits with possible values 0, 1, 2 or 3).
130 :param scoreType: The default HARRIS_SCORE means that Harris algorithm is used to rank features (the score is written to ``KeyPoint::score`` and is used to retain best ``nfeatures`` features); FAST_SCORE is alternative value of the parameter that produces slightly less stable keypoints, but it is a little faster to compute.
132 :param patchSize: size of the patch used by the oriented BRIEF descriptor. Of course, on smaller pyramid layers the perceived image area covered by a feature will be larger.
136 Finds keypoints in an image and computes their descriptors
138 .. ocv:function:: void ORB::operator()(InputArray image, InputArray mask, vector<KeyPoint>& keypoints, OutputArray descriptors, bool useProvidedKeypoints=false ) const
140 .. ocv:pyfunction:: cv2.ORB.detect(image[, mask]) -> keypoints
141 .. ocv:pyfunction:: cv2.ORB.compute(image, keypoints[, descriptors]) -> keypoints, descriptors
142 .. ocv:pyfunction:: cv2.ORB.detectAndCompute(image, mask[, descriptors[, useProvidedKeypoints]]) -> keypoints, descriptors
145 :param image: The input 8-bit grayscale image.
147 :param mask: The operation mask.
149 :param keypoints: The output vector of keypoints.
151 :param descriptors: The output descriptors. Pass ``cv::noArray()`` if you do not need it.
153 :param useProvidedKeypoints: If it is true, then the method will use the provided vector of keypoints instead of detecting them.
158 .. ocv:class:: BRISK : public Feature2D
160 Class implementing the BRISK keypoint detector and descriptor extractor, described in [LCS11]_.
162 .. [LCS11] Stefan Leutenegger, Margarita Chli and Roland Siegwart: BRISK: Binary Robust Invariant Scalable Keypoints. ICCV 2011: 2548-2555.
166 The BRISK constructor
168 .. ocv:function:: BRISK::BRISK(int thresh=30, int octaves=3, float patternScale=1.0f)
170 .. ocv:pyfunction:: cv2.BRISK([, thresh[, octaves[, patternScale]]]) -> <BRISK object>
172 :param thresh: FAST/AGAST detection threshold score.
174 :param octaves: detection octaves. Use 0 to do single scale.
176 :param patternScale: apply this scale to the pattern used for sampling the neighbourhood of a keypoint.
180 The BRISK constructor for a custom pattern
182 .. ocv:function:: BRISK::BRISK(std::vector<float> &radiusList, std::vector<int> &numberList, float dMax=5.85f, float dMin=8.2f, std::vector<int> indexChange=std::vector<int>())
184 .. ocv:pyfunction:: cv2.BRISK(radiusList, numberList[, dMax[, dMin[, indexChange]]]) -> <BRISK object>
186 :param radiusList: defines the radii (in pixels) where the samples around a keypoint are taken (for keypoint scale 1).
188 :param numberList: defines the number of sampling points on the sampling circle. Must be the same size as radiusList..
190 :param dMax: threshold for the short pairings used for descriptor formation (in pixels for keypoint scale 1).
192 :param dMin: threshold for the long pairings used for orientation determination (in pixels for keypoint scale 1).
194 :param indexChanges: index remapping of the bits.
198 Finds keypoints in an image and computes their descriptors
200 .. ocv:function:: void BRISK::operator()(InputArray image, InputArray mask, vector<KeyPoint>& keypoints, OutputArray descriptors, bool useProvidedKeypoints=false ) const
202 .. ocv:pyfunction:: cv2.BRISK.detect(image[, mask]) -> keypoints
203 .. ocv:pyfunction:: cv2.BRISK.compute(image, keypoints[, descriptors]) -> keypoints, descriptors
204 .. ocv:pyfunction:: cv2.BRISK.detectAndCompute(image, mask[, descriptors[, useProvidedKeypoints]]) -> keypoints, descriptors
206 :param image: The input 8-bit grayscale image.
208 :param mask: The operation mask.
210 :param keypoints: The output vector of keypoints.
212 :param descriptors: The output descriptors. Pass ``cv::noArray()`` if you do not need it.
214 :param useProvidedKeypoints: If it is true, then the method will use the provided vector of keypoints instead of detecting them.
218 .. ocv:class:: FREAK : public DescriptorExtractor
220 Class implementing the FREAK (*Fast Retina Keypoint*) keypoint descriptor, described in [AOV12]_. The algorithm propose a novel keypoint descriptor inspired by the human visual system and more precisely the retina, coined Fast Retina Key- point (FREAK). A cascade of binary strings is computed by efficiently comparing image intensities over a retinal sampling pattern. FREAKs are in general faster to compute with lower memory load and also more robust than SIFT, SURF or BRISK. They are competitive alternatives to existing keypoints in particular for embedded applications.
222 .. [AOV12] A. Alahi, R. Ortiz, and P. Vandergheynst. FREAK: Fast Retina Keypoint. In IEEE Conference on Computer Vision and Pattern Recognition, 2012. CVPR 2012 Open Source Award Winner.
226 * An example on how to use the FREAK descriptor can be found at opencv_source_code/samples/cpp/freak_demo.cpp
230 The FREAK constructor
232 .. ocv:function:: FREAK::FREAK( bool orientationNormalized=true, bool scaleNormalized=true, float patternScale=22.0f, int nOctaves=4, const vector<int>& selectedPairs=vector<int>() )
234 :param orientationNormalized: Enable orientation normalization.
235 :param scaleNormalized: Enable scale normalization.
236 :param patternScale: Scaling of the description pattern.
237 :param nOctaves: Number of octaves covered by the detected keypoints.
238 :param selectedPairs: (Optional) user defined selected pairs indexes,
242 Select the 512 best description pair indexes from an input (grayscale) image set. FREAK is available with a set of pairs learned off-line. Researchers can run a training process to learn their own set of pair. For more details read section 4.2 in: A. Alahi, R. Ortiz, and P. Vandergheynst. FREAK: Fast Retina Keypoint. In IEEE Conference on Computer Vision and Pattern Recognition, 2012.
244 We notice that for keypoint matching applications, image content has little effect on the selected pairs unless very specific what does matter is the detector type (blobs, corners,...) and the options used (scale/rotation invariance,...). Reduce corrThresh if not enough pairs are selected (43 points --> 903 possible pairs)
246 .. ocv:function:: vector<int> FREAK::selectPairs(const vector<Mat>& images, vector<vector<KeyPoint> >& keypoints, const double corrThresh = 0.7, bool verbose = true)
248 :param images: Grayscale image input set.
249 :param keypoints: Set of detected keypoints
250 :param corrThresh: Correlation threshold.
251 :param verbose: Prints pair selection informations.
255 .. ocv:class:: KAZE : public Feature2D
257 Class implementing the KAZE keypoint detector and descriptor extractor, described in [ABD12]_. ::
259 class CV_EXPORTS_W KAZE : public Feature2D
263 CV_WRAP explicit KAZE(bool extended, bool upright, float threshold = 0.001f,
264 int octaves = 4, int sublevels = 4, int diffusivity = DIFF_PM_G2);
267 .. note:: AKAZE descriptor can only be used with KAZE or AKAZE keypoints
269 .. [ABD12] KAZE Features. Pablo F. Alcantarilla, Adrien Bartoli and Andrew J. Davison. In European Conference on Computer Vision (ECCV), Fiorenze, Italy, October 2012.
275 .. ocv:function:: KAZE::KAZE(bool extended, bool upright, float threshold, int octaves, int sublevels, int diffusivity)
277 :param extended: Set to enable extraction of extended (128-byte) descriptor.
278 :param upright: Set to enable use of upright descriptors (non rotation-invariant).
279 :param threshold: Detector response threshold to accept point
280 :param octaves: Maximum octave evolution of the image
281 :param sublevels: Default number of sublevels per scale level
282 :param diffusivity: Diffusivity type. DIFF_PM_G1, DIFF_PM_G2, DIFF_WEICKERT or DIFF_CHARBONNIER
286 .. ocv:class:: AKAZE : public Feature2D
288 Class implementing the AKAZE keypoint detector and descriptor extractor, described in [ANB13]_. ::
290 class CV_EXPORTS_W AKAZE : public Feature2D
294 CV_WRAP explicit AKAZE(int descriptor_type, int descriptor_size = 0, int descriptor_channels = 3,
295 float threshold = 0.001f, int octaves = 4, int sublevels = 4, int diffusivity = DIFF_PM_G2);
298 .. note:: AKAZE descriptors can only be used with KAZE or AKAZE keypoints. Try to avoid using *extract* and *detect* instead of *operator()* due to performance reasons.
300 .. [ANB13] Fast Explicit Diffusion for Accelerated Features in Nonlinear Scale Spaces. Pablo F. Alcantarilla, Jesús Nuevo and Adrien Bartoli. In British Machine Vision Conference (BMVC), Bristol, UK, September 2013.
304 The AKAZE constructor
306 .. ocv:function:: AKAZE::AKAZE(int descriptor_type, int descriptor_size, int descriptor_channels, float threshold, int octaves, int sublevels, int diffusivity)
308 :param descriptor_type: Type of the extracted descriptor: DESCRIPTOR_KAZE, DESCRIPTOR_KAZE_UPRIGHT, DESCRIPTOR_MLDB or DESCRIPTOR_MLDB_UPRIGHT.
309 :param descriptor_size: Size of the descriptor in bits. 0 -> Full size
310 :param descriptor_channels: Number of channels in the descriptor (1, 2, 3)
311 :param threshold: Detector response threshold to accept point
312 :param octaves: Maximum octave evolution of the image
313 :param sublevels: Default number of sublevels per scale level
314 :param diffusivity: Diffusivity type. DIFF_PM_G1, DIFF_PM_G2, DIFF_WEICKERT or DIFF_CHARBONNIER