Github : https://github.com/bboz/Image_Processing
1.Hafta Ders İçeriği
import matplotlib.pyplot as plt
#import matplotlib.image as mpimg
#img_mpimg.imread('test1.jpg')
#%matplotlib inline
import numpy as np
im_1=plt.imread("1.jpg")
plt.imshow(im_1)
plt.show()
def fonk1(image_1):
print("Resmin boyutu = ",image_1.ndim,"\n")
print("Resmin Shape değeri = ",image_1.shape,"\n")
print("Red için min değer = ",image_1[:,:,0].min(),"\n")
#':' hangi yerde kullanıldıysa oranın tamamından bahsediyor örneğin burada tüm satırlar ve sütünlar
print("Red için max değer = ",image_1[:,:,0].max(),"\n")
# şuan kırmızı için olan değerlere bakıyoruz 1 olsa yeşil 2 olursa mavi
fonk1(im_1)
im_1[:,:,0]=im_1[:,:,0]+50
plt.imshow(im_1)
plt.show()
def my_function_1(my_img):
print("eksen sayisi : ",my_img.ndim)
print("eksen degerleri :",my_img.shape)
print("en kucuk kirmizi renk degeri : ",np.min(my_img[:,:,0]))
print("en kucuk kirmizi renk degeri : ",np.max(my_img[:,:,0]))
print("en kucuk yesil renk degeri : ",np.min(my_img[:,:,1]))
print("en kucuk yesil renk degeri : ",np.max(my_img[:,:,1]))
print("en kucuk mavi renk degeri : ",np.min(my_img[:,:,2]))
print("en kucuk mavi renk degeri : ",np.max(my_img[:,:,2]))
my_function_1(im_1)
2.Hafta Ders İçeriği
import os
cwd = os.getcwd()
cwd
os.chdir("C:\\imageProcessing2018")
import matplotlib.pyplot as plt
import numpy as np
im_1=plt.imread("1.jpg")
plt.imshow(im_1)
plt.show()
from scipy import stats
from scipy.stats import iqr
from scipy.stats import skew
#mean=sum(img)/len(img)
#plt.subplot(1,2,2),plt.imshow(255-img)
im_1.setflags(write=1)
# resim sadece read-only özelliğine sahipti üzerinde değişiklik yapabilmek için değiştirdim
im_1.flags
im_1[:,:,0]=im_1[:,:,0]+150
# resim üzerinde kırmızı tonu için renk değişimi
plt.imshow(im_1)
plt.show()
def my_function(x):
return 255-x
def inverse(image):
image[:,:,0]=my_function(image[:,:,0])
image[:,:,1]=my_function(image[:,:,1])
image[:,:,2]=my_function(image[:,:,2])
def mean(image):
print("Kirmizi icin renk ortalamasi : ",np.mean(image[:,:,0]))
print("Yesil icin renk ortalamasi : ",np.mean(image[:,:,1]))
print("Mavi icin renk ortalamasi : ",np.mean(image[:,:,2]))
def median(image):
print("Kirmizi icin renk ortanca degeri : ",np.median(image[:,:,0]))
print("Yesil icin renk ortanca degeri : ",np.median(image[:,:,1]))
print("Mavi icin renk ortanca degeri : ",np.median(image[:,:,2]))
def mode(image):
print("Kirmizi icin renk modu : ",stats.mode(image[:,:,0]))
print("Yesil icin renk modu : ",stats.mode(image[:,:,1]))
print("Mavi icin renk modu : ",stats.mode(image[:,:,2]))
def my_H(image):
H={}
for i in range (image.shape[0]):
for j in range (image.shape[1]):
if(image[i,j,0] in H.keys()):
H[image[i,j,0]]=H[image[i,j,0]]+1
else:
H[image[i,j,0]]=1
if(image[i,j,1] in H.keys()):
H[image[i,j,1]]=H[image[i,j,1]]+1
else:
H[image[i,j,1]]=1
if(image[i,j,2] in H.keys()):
H[image[i,j,2]]=H[image[i,j,2]]+1
else:
H[image[i,j,2]]=1
plt.bar(list(H.keys()),list( H.values()), color='r')
plt.show()
def ceyreklik(image):
print("Kirmizi icin Q1 degeri = ",np.percentile(image[:,:,0],25))
print("Kirmizi icin Q2 degeri = ",np.percentile(image[:,:,0],50))
print("Kirmizi icin Q3 degeri = ",np.percentile(image[:,:,0],75))
print("Yesil icin Q1 degeri = ",np.percentile(image[:,:,1],25))
print("Yesil icin Q2 degeri = ",np.percentile(image[:,:,1],50))
print("Yesil icin Q3 degeri = ",np.percentile(image[:,:,1],75))
print("Mavi icin Q1 degeri = ",np.percentile(image[:,:,2],25))
print("Mavi icin Q2 degeri = ",np.percentile(image[:,:,2],50))
print("Mavi icin Q3 degeri = ",np.percentile(image[:,:,2],75))
print("Kirmizi icin iqr degeri = ",iqr(image[:,:,0]))
print("Yesil icin iqr degeri = ",iqr(image[:,:,1]))
print("Mavi icin iqr degeri = ",iqr(image[:,:,2]))
def ss(image):
print("Kirmizi icin skewness degeri = ",skew(image[:,:,0]))
print("Yesil icin skewness degeri = ",skew(image[:,:,1]))
print("Mavi icin skewness degeri = ",skew(image[:,:,2]))
def aralik(image):
print("Kirmizi icin range araligi = ",image[:,:,0].max()-image[:,:,0].min())
print("Yesil icin range araligi = ",image[:,:,0].max()-image[:,:,1].min())
print("Mavi icin range araligi = ",image[:,:,0].max()-image[:,:,2].min())
inverse(im_1)
mean(im_1)
median(im_1)
mode(im_1)
plt.imshow(im_1)
plt.show()
my_H(im_1)
ceyreklik(im_1)
ss(im_1)
aralik(im_1)
3.Hafta Ders İçeriği
#!/usr/bin/env python
# coding: utf-8
import numpy as np
import os
cwd = os.getcwd()
cwd
os.chdir("C:\\imageProcessing2018")
import matplotlib.pyplot as plt
import numpy as np
im1=plt.imread("1.jpg")
plt.imshow(im1)
plt.show()
im1.setflags(write=1)
# resim sadece read-only özelliğine sahipti üzerinde değişiklik yapabilmek için değiştirdim
get_ipython().run_line_magic('matplotlib', 'inline')
def get_distance(v,w=[1/3,1/3,1/3]): #w ağırlık değeri
a,b,c=v[0],v[1],v[2]
w1,w2,w3=w[0],w[1],w[2]
#d=((a*w1)**2+(b*w2)**2+(c*w3)**2)**.5 #sqrt işlemi var
d=((a**2)*w1+(b**2)*w2+(c**2)*w3)**.5
return d
my_RGB=[1,2,3]
gray_level=get_distance(my_RGB)
print(gray_level)
my_RGB=[10,20,3]
gray_level=get_distance(my_RGB,[.6,.3,.1])
print(gray_level)
def convert_rgb_to_gray_level(im_1):
m=im_1.shape[0]
n=im_1.shape[1]
im_2=np.zeros((m,n))
for i in range(m):
for j in range(n):
im_2[i,j]=get_distance(im_1[i,j,:])
return im_2
im2=convert_rgb_to_gray_level(im1)
type(im2)
im1.shape,im2.shape
plt.imshow(im2,cmap='gray')
plt.show()
def convert_gray_to_BW(image_gray_level):
m=image_gray_level.shape[0]
n=image_gray_level.shape[1]
im_bw=np.zeros((m,n))
for i in range(m):
for j in range(n):
if image_gray_level[i,j]>120:
im_bw[i,j]=1
else:
im_bw[i,j]=0
return im_bw
im3=convert_gray_to_BW(im2)
plt.subplot(1,3,1),plt.imshow(im1)
plt.subplot(1,3,2),plt.imshow(im2,cmap='gray')
plt.subplot(1,3,3),plt.imshow(im3,cmap='gray')
plt.imshow(im3,cmap='gray')
plt.show()
 |
| RGB |
 |
| Gray |
 |
| BW |
4.Hafta Ders İçeriği
# coding: utf-8
import numpy as np
#resim üzerinden bir parça alıp onu mask ile çarpabilir toplayabilir herhangibir işleme tabi tutabiliriz.
import matplotlib.pyplot as plt
mask_0=np.array([1,1,1,1,1,1,1,1,1]).reshape(3,3)
mask_0=(mask_0)/9
mask_1=np.random.randint(5,size=9).reshape(3,3)
mask_2=np.random.randint(5,size=9).reshape(3,3)
print(mask_1)
print("------------")
print(mask_2)
print("------------")
print(mask_1*mask_2)
print("------------")
print(mask_1*mask_0)
def get_distance(v,w=[1/3,1/3,1/3]): #w ağırlık değeri
a,b,c=v[0],v[1],v[2]
w1,w2,w3=w[0],w[1],w[2]
#d=((a*w1)**2+(b*w2)**2+(c*w3)**2)**.5 #sqrt işlemi var
d=((a**2)*w1+(b**2)*w2+(c**2)*w3)**.5
return d
def convert_rgb_to_gray_level(im_1):
m=im_1.shape[0]
n=im_1.shape[1]
im_2=np.zeros((m,n))
for i in range(m):
for j in range(n):
im_2[i,j]=get_distance(im_1[i,j,:])
return im_2
sum(sum(mask_1*mask_0))
def get_default_mask_for_mean():
return np.array([1,1,1,1,1,1,1,1,1]).reshape(3,3)/9
def apply_mask(part_of_image):
mask=get_default_mask_for_mean()
return np.average(part_of_image)
im1=plt.imread("1.jpg") #ortalama ile yapılan örnek
im1=convert_rgb_to_gray_level(im1)
plt.imshow(im1,cmap='gray')
plt.show()
m=im1.shape[0]
n=im1.shape[1]
im2=np.zeros((m,n))
for i in range(1,m):
for j in range(1,n):
poi=im1[i-1:i+2,j-1:j+2]
im2[i,j]=apply_mask(poi) #part of image = poi
plt.imshow(im2,cmap='gray')
plt.show()
mask_1=np.random.randint(5,size=9).reshape(3,3)
mask_1
mask_1[:,0:1]
im2[3:6,10:13]
s_1=im2[3:6,10:13].reshape(9)
s_1.sort()
s_1[5]
def get_median(poi):
s_1=poi.reshape(1,9)
s_1.sort()
return s_1[0,4]
im1=plt.imread("1.jpg") # ortanca ile yapılan örnek
im1=convert_rgb_to_gray_level(im1)
plt.imshow(im1,cmap='gray')
plt.show()
m=im1.shape[0]
n=im1.shape[1]
im2=np.zeros((m,n))
for i in range(1,m-1):
for j in range(1,n-1):
poi=im1[i-1:i+2,j-1:j+2]
im2[i,j]=get_median(poi) #part of image = poi
plt.imshow(im2,cmap='gray')
plt.show()
 |
| Mask Uygulanmamış Resim |
 |
| Mask Uygulanmış Resim |
5.Hafta Ders İçeriği
#!/usr/bin/env python
# coding: utf-8
import numpy as np #resmimizi gray level daha sonra bw aktarıcaz t için e - 6 t için i - 2
import matplotlib.pyplot as plt
im1=plt.imread("3.png")
plt.imshow(im1)
plt.show()
def get_distance(v,w=[1/3,1/3,1/3]): #w ağırlık değeri
a,b,c=v[0],v[1],v[2]
w1,w2,w3=w[0],w[1],w[2]
#d=((a*w1)**2+(b*w2)**2+(c*w3)**2)**.5 #sqrt işlemi var
d=((a**2)*w1+(b**2)*w2+(c**2)*w3)**.5
return d
def convert_rgb_to_gray_level(im_1):
m=im_1.shape[0]
n=im_1.shape[1]
im_2=np.zeros((m,n))
for i in range(m):
for j in range(n):
im_2[i,j]=get_distance(im_1[i,j,:])
return im_2
def convert_rgb_to_BW(image_gray_level):
m=image_gray_level.shape[0]
n=image_gray_level.shape[1]
im_bw=np.zeros((m,n))
for i in range(m):
for j in range(n):
if image_gray_level[i,j]==0:
im_bw[i,j]=0
else:
im_bw[i,j]=1
return im_bw
im2=convert_rgb_to_gray_level(im1)
plt.imshow(im2,cmap='gray')
plt.show()
im3=convert_rgb_to_BW(im2)
plt.imshow(im3,cmap='gray')
plt.show()
def pixel_compenent(resim): #evde pixel compenenti ayrı fonksiyon yap ayır daha güzel olur
m=resim.shape[0] # siyah 0 beyaz 1
n=resim.shape[1]
external=0
internal=0
for i in range (1,m-1):
for j in range(1,n-1):
poi=resim[i:i+2,j:j+2]
siyah=0
beyaz=0
for k in range(2):
for l in range(2):
if poi[k][l]==1:
beyaz=beyaz+1
else:
siyah=siyah+1
if(siyah>beyaz and beyaz>0):
internal=internal+1
elif(beyaz>siyah and siyah>0):
external=external+1
print("dış kenar ",external)
print("iç kenar ",internal)
print("Nesne sayisi",(external-internal)/4)
pixel_compenent(im3)
6.Hafta Ders İçeriği
import numpy as np
import matplotlib.pyplot as plt
import random as random
def my_create_m():
return np.random.randint(0,2,(28,28))
def MBR_create_28_by_28_with_0_1(matris_a):
m=matris_a.shape[0]
n=matris_a.shape[1]
x_min=m
x_max=0 #başlangıç değerleri olası en olumsuz durum
y_min=n
y_max=0
for i in range(m):
for j in range(n):
if (matris_a[i,j]==1 and x_min>i): # resim/matris üzerinden
x_min=i # x_min, .... güncelleniyor
if (matris_a[i,j]==1 and x_maxj):
y_min=j
if (matris_a[i,j]==1 and y_maxbenzerlik_max):
benzerlik_max=get_similarity(a,b)
c=b
print("En Yüksek Benzerlik : ",benzerlik_max)
plt.imshow(a,interpolation='nearest',cmap='gray')
plt.show()
plt.imshow(c,interpolation='nearest',cmap='gray')
plt.show()
get_similarity_for_100_characters()
7.Hafta (Ödev)
import numpy as np
import matplotlib.pyplot as plt
im1=plt.imread("7.jpg")
im1.setflags(write="1")
plt.imshow(im1)
plt.show()
def get_distance(v,w=[1/3,1/3,1/3]): #w ağırlık değeri
a,b,c=v[0],v[1],v[2]
w1,w2,w3=w[0],w[1],w[2]
#d=((a*w1)**2+(b*w2)**2+(c*w3)**2)**.5 #sqrt işlemi var
d=((a**2)*w1+(b**2)*w2+(c**2)*w3)**.5
return d
def convert_rgb_to_gray_level(im_1):
m=im_1.shape[0]
n=im_1.shape[1]
im_2=np.zeros((m,n))
for i in range(m):
for j in range(n):
im_2[i,j]=get_distance(im_1[i,j,:])
return im_2
def convert_rgb_to_BW(image_gray_level):
m=image_gray_level.shape[0]
n=image_gray_level.shape[1]
im_bw=np.zeros((m,n))
for i in range(m):
for j in range(n):
if image_gray_level[i,j]>120:
im_bw[i,j]=0
else:
im_bw[i,j]=1
return im_bw
def flatlist(liste):
flat=[]
for i in range (len(liste)):
for j in range (len(liste)):
flat.append(liste[i][j])
return flat
def apply_mask(liste,mask=[1,1,1,1,1,1,1,1,1]):
liste2=flatlist(liste)
result=0
if liste2==mask:
result=1
return result
def apply_mask2(liste,mask=[0,0,0,0,0,0,0,0,0]):
liste2=flatlist(liste)
result=1
if liste2==mask:
result=0
return result
im2=convert_rgb_to_gray_level(im1)
im3=convert_rgb_to_BW(im2)
def eresion(resim):
m=resim.shape[0] # siyah 0 beyaz 1
n=resim.shape[1]
im4=np.zeros((m,n))
for i in range (1,m-2):
for j in range(1,n-2):
b=resim[i-1:i+2,j-1:j+2]
im4[i,j]=apply_mask(b)
plt.imshow(im4,cmap='gray')
plt.show()
return im4
def dilation(resim):
m=resim.shape[0] # siyah 0 beyaz 1
n=resim.shape[1]
im6=np.zeros((m,n))
for i in range (1,m-2):
for j in range(1,n-2):
b=resim[i-1:i+2,j-1:j+2]
im6[i,j]=apply_mask2(b)
plt.imshow(im6,cmap='gray')
plt.show()
return im6
im5=eresion(im3)
plt.imshow(im3,cmap='gray')
plt.show()
print(im5.shape,im3.shape)
im7=dilation(im3)
plt.imshow(im3,cmap='gray')
plt.show()
print(im7.shape,im3.shape)
im10=plt.imread("1.jpg")
im10.setflags(write="1")
plt.imshow(im10)
plt.show()
im11=convert_rgb_to_gray_level(im10)
plt.imshow(im11,cmap='gray')
plt.show()
im12=convert_rgb_to_BW(im11)
plt.imshow(im12,cmap='gray')
plt.show()
 |
| Orjinal Resim |
 |
| Eresion Uygulanmış Hali |
 |
| Dilation Uygulanmış Hali |
10.Hafta Ders İçeriği
#importing the libraries
import numpy as np
import matplotlib.pyplot as plt
from sklearn import datasets
from sklearn.decomposition import PCA
import matplotlib.pyplot as plt
from mpl_toolkits import mplot3d
iris =datasets.load_iris()
setosa=iris.data[0:50]
versicolor=iris.data[50:100]
virginica=iris.data[100:150]
setosa_sepal_length_avg=np.mean(setosa[:,0])
setosa_sepal_width_avg=np.mean(setosa[:,1])
setosa_petal_length_avg=np.mean(setosa[:,2])
setosa_petal_width_avg=np.mean(setosa[:,3])
versicolor_sepal_length_avg=np.mean(versicolor[:,0])
versicolor_sepal_width_avg=np.mean(versicolor[:,1])
versicolor_petal_length_avg=np.mean(versicolor[:,2])
versicolor_petal_width_avg=np.mean(versicolor[:,3])
virginica_sepal_length_avg=np.mean(virginica[:,0])
virginica_sepal_width_avg=np.mean(virginica[:,1])
virginica_petal_length_avg=np.mean(virginica[:,2])
virginica_petal_width_avg=np.mean(virginica[:,3])
#eğer veriler sınıflara ayrılmışsa elimizdeki veri setinde sınıf sayısını bulabiliriz.
def sinifsayisibul():
h=[]
for i in range(len(iris.target)):
if (iris.target[i] in h):
continue
else:
h.append(iris.target[i])
return len(h)
sinifsayisibul()
fig=plt.figure()
ax=plt.axes(projection='3d')
xdata=iris.data[:,0]
ydata=iris.data[:,2]
zdata=iris.data[:,3]
label=iris.target
ax.scatter3D(xdata[100:149], ydata[100:149], zdata[100:149],marker='o')
ax.scatter3D(xdata[50:99], ydata[50:99], zdata[50:99],marker='x')
ax.scatter3D(xdata[0:49], ydata[0:49], zdata[0:49],marker='x')
def get_mu_s():
mu_0=[5,2,0]
mu_1=[4,4,0]
mu_2=[3,2,5]
return mu_0,mu_1,mu_2
def get_distance(mu,point):
x=mu[0]-point[0]
y=mu[1]-point[1]
z=mu[2]-point[2]
return (x**2+y**2+z**2)**0.5
def get_class_for_one_instance(flower):
mu_s = get_mu_s()
d_0 = get_distance(mu_s[0], flower)
d_1 = get_distance(mu_s[1], flower)
d_2 = get_distance(mu_s[2], flower)
if((d_0 < d_1) and (d_0 < d_2)):
return "0"
if((d_1 < d_0) and (d_1 < d_2)):
return "1"
if((d_2 < d_1) and (d_2 < d_0)):
return "2"
def my_f_1(s_1=125):
x=iris.data[s_1][0]
y=iris.data[s_1][2]
z=iris.data[s_1][3]
my_f_1=[x,y,z]
r=get_class_for_one_instance(my_f_1)
print(r)
for i in range(150):
my_f_1(i)
def get_flower(i):
#iris data üzerinde bana i değerinde gelen datanın 0 2 3 sütunun donduruyorum
x=iris.data[i][0]
y=iris.data[i][2]
z=iris.data[i][3]
return [x,y,z]
def update_mu():
#150 ya kadar her bir çicek için bana class bilgisi ver güncelle ve geriye mu'leri ver
hata="yok"
mu_0_counter=0.0001
mu_0_sum=0
mu_1_counter=0.0001
mu_1_sum=0
mu_2_counter=0.0001
mu_2_sum=0
c_1=[]
c_2=[]
c_3=[]
for i in range(150):
my_flower_data=get_flower(i)
f_class = get_class_for_one_instance(my_flower_data)
hata="var"
if(f_class=="0"):
c_1.append(my_flower_data)
if(f_class=="1"):
c_2.append(my_flower_data)
if(f_class=="2"):
c_3.append(my_flower_data)
return c_1,c_2,c_3
c_1,c_2,c_3 = update_mu()
print(len(c_1))
print(len(c_2))
print(len(c_3))
 |
| K-Means Sınıflandırma |
Hiç yorum yok:
Yorum Gönder