def matchImages(original, image_to_compare, useFlann=False):
if useRootSIFT:
# extract RootSIFT descriptors
#print('Using RootSIFT')
kps = None
rs = RootSIFT()
kp_1, desc_1 = rs.compute(original, kps)
kp_2, desc_2 = rs.compute(image_to_compare, kps)
else:
sift = cv2.xfeatures2d.SIFT_create()
kp_1, desc_1 = sift.detectAndCompute(original, None)
kp_2, desc_2 = sift.detectAndCompute(image_to_compare, None)
if useFlann:
FLANN_INDEX_KDTREE = 0
index_params = dict(algorithm=FLANN_INDEX_KDTREE, trees=5)
search_params = dict(checks=50)
print('Using Flann Matcher')
matcher = cv2.FlannBasedMatcher(index_params, search_params)
else:
print('Using Brute Force Matcher')
matcher = cv2.DescriptorMatcher_create("BruteForce")
matches = matcher.knnMatch(desc_1, desc_2, k=2)
good_points = []
for m, n in matches:
if m.distance < LRatio * n.distance:
good_points.append(m)
print(len(good_points))
result = cv2.drawMatches(original, kp_1, image_to_compare, kp_2,
good_points, None)
# cv2.imshow("result", cv2.resize(result,(800,600)))
return (good_points, result)
def getRootSIFT(gray):
sift = cv2.xfeatures2d.SIFT_create()
kp, des = sift.detectAndCompute(gray, None)
# extract RootSIFT descriptors
rs = RootSIFT()
kp, des = rs.compute(gray, kp)
return kp, des
def detectAndDescribe(self, image, useRootSIFT=False):
# convert the image to grayscale
# gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
# check to see if we are using OpenCV 3.X
if self.isv3:
if useRootSIFT:
# extract RootSIFT descriptors
# print('Using RootSIFT')
kps = None
rs = RootSIFT()
(kps, features) = rs.compute(image, kps)
else:
# detect and extract features from the image
descriptor = cv2.xfeatures2d.SIFT_create()
# descriptor = cv2.xfeatures2d.SURF_create()
# descriptor = cv2.ORB_create()
(kps, features) = descriptor.detectAndCompute(image, None)
# otherwise, we are using OpenCV 2.4.X
else:
# detect keypoints in the image
detector = cv2.FeatureDetector_create("SIFT")
kps = detector.detect(image)
# extract features from the image
extractor = cv2.DescriptorExtractor_create("SIFT")
(kps, features) = extractor.compute(image, kps)
# convert the keypoints from KeyPoint objects to NumPy
# arrays
kps = np.float32([kp.pt for kp in kps])
# return a tuple of keypoints and features
return (kps, features)
def get_local_features(params, image):
# detect Difference of Gaussian keypoints in the image
detector = cv2.FeatureDetector_create(params["keypoint_type"])
kps = detector.detect(image)
# extract RootSIFT descriptors
rs = RootSIFT()
(kps, descs) = rs.compute(image, kps)
# Devolvemos los descriptores para una imagen
return descs
from __future__ import print_function
from rootsift import RootSIFT
import argparse
import cv2
import imutils
# construct the argument parser and parse the arguments
ap = argparse.ArgumentParser()
ap.add_argument("-i", "--image", required=True, help="Path to the image")
args = vars(ap.parse_args())
# load the input image, convert it to grayscale
image = cv2.imread(args["image"])
gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
# extract local invariant descriptors
extractor = RootSIFT()
(kps, descs) = extractor.compute(gray, None)
# show the shape of the keypoints and local invariant descriptors array
print("[INFO] # of keypoints detected: {}".format(len(kps)))
print("[INFO] feature vector shape: {}".format(descs.shape))
# it to grayscale
image = cv2.imread("all_souls_000035.jpg")
gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
# detect Difference of Gaussian keypoints in the image
detector = cv2.FeatureDetector_create("SIFT")
kps = detector.detect(image)
# extract normal SIFT descriptors
extractor = cv2.DescriptorExtractor_create("SIFT")
(kps_sift, descs_sift) = extractor.compute(image, kps)
print "SIFT: kps=%d, descriptors=%s " % (len(kps_sift), descs_sift.shape)
# extract RootSIFT descriptors
rs = RootSIFT()
(kps_rootsift, descs_rootsift) = rs.compute(image, kps)
print "RootSIFT: kps=%d, descriptors=%s " % (len(kps_rootsift), descs_rootsift.shape)
pylab.figure()
rgbImage = cv2.cvtColor(image, cv2.COLOR_BGR2RGB)
img_SIFT = cv2.drawKeypoints(rgbImage,kps_sift,None,(255,0,255),4)
pylab.gray()
pylab.subplot(2,1,1)
pylab.imshow(img_SIFT)
pylab.axis('off')
img_rootsift = cv2.drawKeypoints(rgbImage,kps_rootsift,None,(255,0,255),4)
pylab.gray()
pylab.subplot(2,1,2)
pylab.imshow(img_rootsift)
# USAGE
# python driver.py
# import the necessary packages
from rootsift import RootSIFT
import cv2
# load the image we are going to extract descriptors from and convert
# it to grayscale
image = cv2.imread("example.png")
gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
# detect Difference of Gaussian keypoints in the image
detector = cv2.FeatureDetector_create("SIFT")
kps = detector.detect(gray)
# extract normal SIFT descriptors
extractor = cv2.DescriptorExtractor_create("SIFT")
(kps, descs) = extractor.compute(gray, kps)
print("SIFT: kps=%d, descriptors=%s " % (len(kps), descs.shape))
# extract RootSIFT descriptors
rs = RootSIFT()
(kps, descs) = rs.compute(gray, kps)
print("RootSIFT: kps=%d, descriptors=%s " % (len(kps), descs.shape))
# import the necessary packages
from rootsift import RootSIFT
import cv2
# load the image we are going to extract descriptors from and convert
# it to grayscale
image = cv2.imread("example.jpg")
gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
# detect Difference of Gaussian keypoints in the image
detector = cv2.FeatureDetector_create("SIFT")
kps = detector.detect(image)
# extract normal SIFT descriptors
extractor = cv2.DescriptorExtractor_create("SIFT")
(kps, descs) = extractor.compute(image, kps)
print "SIFT: kps=%d, descriptors=%s " % (len(kps), descs.shape)
# extract RootSIFT descriptors
rs = RootSIFT()
(kps, descs) = rs.compute(image, kps)
print "RootSIFT: kps=%d, descriptors=%s " % (len(kps), descs.shape)
# import the necessary packages
from rootsift import RootSIFT
import cv2
# load the image we are going to extract descriptors from and convert
# it to grayscale
image = cv2.imread("example.png")
gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
# detect Difference of Gaussian keypoints in the image
detector = cv2.FeatureDetector_create("SIFT")
kps = detector.detect(image)
# extract normal SIFT descriptors
extractor = cv2.DescriptorExtractor_create("SIFT")
(kps, descs) = extractor.compute(image, kps)
print "SIFT: kps=%d, descriptors=%s " % (len(kps), descs.shape)
# extract RootSIFT descriptors
rs = RootSIFT()
(kps, descs) = rs.compute(image, kps)
print "RootSIFT: kps=%d, descriptors=%s " % (len(kps), descs.shape)
# detect Difference of Gaussian keypoints in the image detector = cv2.xfeatures2d.SIFT_create() kps1 = detector.detectAndCompute(gray1, None) kps2 = detector.detectAndCompute(gray2, None) # extract normal SIFT descriptors extractor = cv2.xfeatures2d.SIFT_create() (kps1, descs1) = extractor.detectAndCompute(gray1, None) (kps2, descs2) = extractor.detectAndCompute(gray2, None) print "SIFT: kps=%d, descriptors=%s " % (len(kps1), descs1.shape) print "SIFT: kps=%d, descriptors=%s " % (len(kps2), descs2.shape) # extract RootSIFT descriptors rs = RootSIFT() (kps1, descs1, kps2, descs2) = rs.compute(gray1, kps1, gray2, kps2) print "RootSIFT: kps=%d, descriptors=%s " % (len(kps1), descs1.shape) print "RootSIFT: kps=%d, descriptors=%s " % (len(kps2), descs2.shape) print "========================================================" print "Slika poklapanja:" print "========================================================" # FLANN parameters FLANN_INDEX_KDTREE = 0 index_params = dict(algorithm = FLANN_INDEX_KDTREE, trees = 5) search_params = dict(checks=50) # or pass empty dictionary flann = cv2.FlannBasedMatcher(index_params,search_params) matches = flann.knnMatch(descs1,descs2,k=2)
# extract RootSIFT descriptors
rs = RootSIFT()
#descriptor = cv2.xfeatures2d.SIFT_create()
matcher = cv2.DescriptorMatcher_create("BruteForce")
objKeyPtList = []
objFtrList = []
for obj in object_list:
imageA = cv2.imread(obj)
# cv2.imshow("Image A", imageA)
grayA = cv2.cvtColor(imageA, cv2.COLOR_BGR2GRAY)
# detect and extract features from the image
(kps1, featuresA) = rs.compute(grayA, None)
# (kps1, featuresA) = descriptor.detectAndCompute(grayA, None)
if type(featuresA) != type(None):
kpsA = np.float32([kp.pt for kp in kps1])
objKeyPtList.append(kpsA)
objFtrList.append(featuresA)
print('Total objects found:', len(objFtrList))
# grab a reference to the video file
vs = cv2.VideoCapture(args["video"])
# allow the camera or video file to warm up
time.sleep(2.0)
def readFromFile(testImagePath):
imgTest = cv2.imread(testImagePath)
imgPath = 'data_images'
testImagePath = 'test_images'
detector = cv2.xfeatures2d.SIFT_create() #inicijalizacija detektora za kljucne tacke
extractor = cv2.xfeatures2d.SIFT_create() #inicijalizacija ekstraktora preko koga cemo uzimati osobine slike - deskriptor
rs = RootSIFT()
#======================================================================
#funkcija koja nalazi broj pozitivnih preklapanja kljucnih tacaka
def findPosMatches(dscTest, dscData):
pozPoklapanja = []
# FLANN parameters
FLANN_INDEX_KDTREE = 0
index_params = dict(algorithm = FLANN_INDEX_KDTREE, trees = 5)
search_params = dict(checks = 100)
flann = cv2.FlannBasedMatcher(index_params,search_params)
matches = flann.knnMatch(dscTest,dscData,k=2)
# Need to draw only good matches, so create a mask
matchesMask = [[0,0] for i in xrange(len(matches))]
#pozPoklapanja = []
# ratio test as per Lowe's paper
for i,(m,n) in enumerate(matches):
if m.distance < 0.7*n.distance:
matchesMask[i]=[1,0]
pozPoklapanja.append(len(matchesMask[i]))
x = len(pozPoklapanja)
return x, matches, matchesMask
#======================================================================
#obrada ucitane slike
gray2 = cv2.cvtColor(imgTest, cv2.COLOR_BGR2GRAY)
kptsTest = detector.detectAndCompute(gray2, None)
(kptsTest, dscTest) = extractor.detectAndCompute(gray2, None)
(kptsTest, dscTest) = rs.compute(gray2, kptsTest)
# inicijalizacija osnovnih parametara u koje cu smestati podatke one slike
# koja ima najvise pozitivnih poklapanja kljucnih tacaka
maxBrojPozPoklapanja = 0
kptsDataBest = []
poklapanja = []
slikaData = imgTest
matchesMask = []
#obrada slika iz dataseta
for imgPath in glob.glob(folderPath + '/*.jpg'):
imgData = cv2.imread(imgPath)
gray1 = cv2.cvtColor(imgData, cv2.COLOR_BGR2GRAY)
kptsData = detector.detectAndCompute(gray1, None)
(kptsData, dscData) = extractor.detectAndCompute(gray1, None)
(kptsData, dscData) = rs.compute(gray1, kptsData)
brPozPoklapanjaSlike, matchesData, matchesMaskData = findPosMatches(dscTest, dscData)
if brPozPoklapanjaSlike > maxBrojPozPoklapanja: #ukoliko slika ima najvise pozitivnih poklapanja
maxBrojPozPoklapanja = brPozPoklapanjaSlike #ona postaje glavni kandidat za "pobednika"
poklapanja = matchesData
slikaData = imgData
kptsDataBest = kptsData
matchesMask = matchesMaskData
print "Slika poklapanja:"
draw_params = dict(matchColor = (0,255,0),
singlePointColor = (255,0,0),
matchesMask = matchesMask,
flags = 0)
img3 = cv2.drawMatchesKnn(imgTest,kptsTest,slikaData,kptsDataBest,poklapanja,None,**draw_params)
plt.imshow(img3,),plt.show()
return img3, slikaData
# it to grayscale
image = cv2.imread("all_souls_000035.jpg")
gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
# detect Difference of Gaussian keypoints in the image
detector = cv2.FeatureDetector_create("SIFT")
kps = detector.detect(image)
# extract normal SIFT descriptors
extractor = cv2.DescriptorExtractor_create("SIFT")
(kps_sift, descs_sift) = extractor.compute(image, kps)
print "SIFT: kps=%d, descriptors=%s " % (len(kps_sift), descs_sift.shape)
# extract RootSIFT descriptors
rs = RootSIFT()
(kps_rootsift, descs_rootsift) = rs.compute(image, kps)
print "RootSIFT: kps=%d, descriptors=%s " % (len(kps_rootsift),
descs_rootsift.shape)
pylab.figure()
rgbImage = cv2.cvtColor(image, cv2.COLOR_BGR2RGB)
img_SIFT = cv2.drawKeypoints(rgbImage, kps_sift, None, (255, 0, 255), 4)
pylab.gray()
pylab.subplot(2, 1, 1)
pylab.imshow(img_SIFT)
pylab.axis('off')
img_rootsift = cv2.drawKeypoints(rgbImage, kps_rootsift, None, (255, 0, 255),
4)
pylab.gray()