tynbl.github.io
常用的图像特征
1. 颜色特征
from skimage import data, img_as_float, exposure
# 如果需要使用参数nbins，需要将图像数据从[0, 255]转换到[0, 1]
camera = img_as_float(data.camera())
# 颜色直方图
hist, bin_centers = exposure.histogram(camera, nbins=10)
print(hist)
print(bin_centers)
[51199 8554 6922 8834 31923 45742 82660 23862 1470 978]
[ 0.05 0.15 0.25 0.35 0.45 0.55 0.65 0.75 0.85 0.95]
SIFT 特征 (DAISY特征)
from skimage.feature import daisy
import matplotlib.pyplot as plt
%matplotlib inline
daisy_feat, daisy_img = daisy(camera,step=180, radius=58, rings=2, histograms=6, visualize=True)
#daisy_feat, daisy_img = daisy(camera, visualize=True)
print(daisy_feat.shape)
plt.imshow(daisy_img)
(3, 3, 104)
<matplotlib.image.AxesImage at 0x6315f28>
skimage – HOG 特征
from skimage.feature import hog
import matplotlib.pyplot as plt
%matplotlib inline
hog_feat, hog_img = hog(camera, visualise=True)
print(hog_feat.shape)
plt.imshow(hog_img)
C:\Anaconda2\envs\py35\lib\site-packages\skimage\feature\_hog.py:119: skimage_deprecation: Default value of `block_norm`==`L1` is deprecated and will be changed to `L2-Hys` in v0.15
'be changed to `L2-Hys` in v0.15', skimage_deprecation)
(311364,)
<matplotlib.image.AxesImage at 0x659d5f8>
