def get_hog_feature(roi):
"""
Compute HOG of ROI.\n
:param roi: Region of interest, ndarray\n
:return: HOG feature, ndarray
"""
winSize = (roi.shape[0], roi.shape[1])
blockSize = (8, 8)
blockStride = (1, 1)
cellSize = (8, 8)
nbins = 9
derivAperture = 1
winSigma = 4.0
histogramNormType = 0
L2HysThreshold = 2.0000000000000001e-01
gammaCorrection = 0
nlevels = 64
# Builder descriptor
descriptor = cv2.HOGDescriptor(winSize, blockSize, blockStride, cellSize,
nbins, derivAperture, winSigma,
histogramNormType, L2HysThreshold,
gammaCorrection, nlevels)
# Compute HOG
winStride = (8, 8)
padding = (8, 8)
locations = []
hist = descriptor.compute(roi, winStride, padding,
locations) # Size: (437400, 1)
return hist / np.sum(hist) # Normalization
def obtain_hog_features(img):
hog = cv2.HOGDescriptor()
winStride = (4, 4)
padding = (4, 4)
locations = ((5, 10), )
#h = hog.compute(img,winStride,padding,locations)
h = hog.compute(img)
return h.flatten()
def model_cut(path):
# initialize the HOG descriptor/person detector
hog = cv2.HOGDescriptor()
hog.setSVMDetector(cv2.HOGDescriptor_getDefaultPeopleDetector())
#path=r"./image" #文件路径
filelist = sorted(os.listdir(path), key=lambda x: int(x[:-4])) #该文件夹下的所有文件
status = []
# loop over the image paths
for imagePath in filelist:
if imagePath == '.DS_Store':
continue
else:
# load the image and resize it to (1) reduce detection time
# and (2) improve detection accuracy
image = cv2.imread(path + '/' + imagePath)
image = imutils.resize(image, width=min(400, image.shape[1]))
orig = image.copy()
# detect people in the image
(rects, weights) = hog.detectMultiScale(image,
winStride=(4, 4),
padding=(8, 8),
scale=1.05)
# draw the original bounding boxes
for (x, y, w, h) in rects:
cv2.rectangle(orig, (x, y), (x + w, y + h), (0, 0, 255), 2)
# apply non-maxima suppression to the bounding boxes using a
# fairly large overlap threshold to try to maintain overlapping
# boxes that are still people
rects = np.array([[x, y, x + w, y + h] for (x, y, w, h) in rects])
pick = non_max_suppression(rects, probs=None, overlapThresh=0.65)
# draw the final bounding boxes
for (xA, yA, xB, yB) in pick:
cv2.rectangle(image, (xA, yA), (xB, yB), (0, 255, 0), 2)
# show some information on the number of bounding boxes
filename = imagePath[imagePath.rfind("/") + 1:]
print("[INFO] {}: {} original boxes, {} after suppression".format(
filename, len(rects), len(pick)))
if len(rects) != 0 and len(pick) != 0:
status.append(1)
else:
status.append(0)
# show the output images
# cv2.imshow("Before NMS", orig)
# cv2.moveWindow("trans:"+filename,500,0)
# cv2.imshow("After NMS", image)
# cv2.waitKey(1)
return len(filelist), status
def _calculate_hog(self, image):
winSize = (60, 120)
blockSize = (6, 6)
blockStride = (6, 6)
cellSize = (6, 6)
nbins = 6
derivAperture = 1
winSigma = 4.
histogramNormType = 0
L2HysThreshold = 2.0000000000000001e-01
gammaCorrection = 0
nlevels = 64
hogs = cv2.HOGDescriptor(winSize, blockSize, blockStride, cellSize, nbins, derivAperture, winSigma,
histogramNormType, L2HysThreshold, gammaCorrection, nlevels)
hog_feats = hogs.compute(image)
return hog_feats.reshape(20, 10, nbins)
def get_hog_features(trainset):
features = []
hog = cv2.HOGDescriptor('../hog.xml')
for img in trainset:
img = np.reshape(img, (28, 28))
cv_img = img.astype(np.uint8)
hog_feature = hog.compute(cv_img)
# hog_feature = np.transpose(hog_feature)
features.append(hog_feature)
features = np.array(features)
features = np.reshape(features, (-1, 324))
return features
from imutils.object_detection import non_max_suppression
import numpy as np
import argparse
# Construct the argument parse and parse the arguments
ap = argparse.ArgumentParser() #creating an ArgumentParser object
ap.add_argument("-i",
"--images",
required=True,
help="path to images directory")
args = vars(
ap.parse_args()
) #parse_args() inspects the command line and converts each argument into the appropriate type and invokes the required action
# Initialize the HOG(Histogram of Oriented Gradients) descriptor/person detector
hog = cv2.HOGDescriptor() #creation of the HOG descriptor
hog.setSVMDetector(
cv2.HOGDescriptor_getDefaultPeopleDetector()
) # we set up the Support Vector Machine to be pre-trained for pedestrian detection
# Loop through the images in images directory
for imagePath in paths.list_images(args["images"]):
# load the image and resize it to reduce detection time & improve accuracy
image = cv2.imread(imagePath)
image = imutils.resize(image, width=min(400, image.shape[1]))
orig = image.copy()
# detect persons in the image
rects, weights = hog.detectMultiScale(image,
winStride=(4, 4),
padding=(8, 8),
def hog_extractor(path):
img = cv.imread(path)
hog = cv.HOGDescriptor(winSize,blockSize,blockStride,cellSize,nbins)
hog_feature = hog.compute(img, winStride, padding).reshape((-1,))
return hog_feature
def obtain_hog_features(img):
hog = cv2.HOGDescriptor()
h = hog.compute(img)
return h.flatten()