Re-design psnr.py code and change image names (#592)

* Change some Image File names & re-code the psnr algorithm (conserving both methods). Also Added new example.

* Selected psnr method and reformat some code from arithmetic_analysis

analysis/compression_analysis/psnr.py

@@ -1,38 +1,39 @@
-import numpy as np
+"""
+	Peak signal-to-noise ratio - PSNR - https://en.wikipedia.org/wiki/Peak_signal-to-noise_ratio
+    Soruce: https://tutorials.techonical.com/how-to-calculate-psnr-value-of-two-images-using-python/
+"""
+
 import math
+
 import cv2
+import numpy as np
 
-def Representational(r,g,b):
+def psnr(original, contrast):
-	return (0.299*r+0.287*g+0.114*b)
+    mse = np.mean((original - contrast) ** 2)
+    if mse == 0:
+        return 100
+    PIXEL_MAX = 255.0
+    PSNR = 20 * math.log10(PIXEL_MAX / math.sqrt(mse))
+    return PSNR
 
-def calculate(img):
-	b,g,r = cv2.split(img)
-	pixelAt = Representational(r,g,b)
-	return pixelAt
 
 def main():
 
-	#Loading images (orignal image and compressed image)
+    # Loading images (original image and compressed image)
-	orignal_image = cv2.imread('orignal_image.png',1)
+    original = cv2.imread('original_image.png')
-	compressed_image = cv2.imread('compressed_image.png',1)
+    contrast = cv2.imread('compressed_image.png', 1)
 
-	#Getting image height and width
+    original2 = cv2.imread('PSNR-example-base.png')
-	height,width = orignal_image.shape[:2]
+    contrast2 = cv2.imread('PSNR-example-comp-10.jpg', 1)
 
-	orignalPixelAt = calculate(orignal_image)
+    # Value expected: 29.73dB
-	compressedPixelAt = calculate(compressed_image)
+    print("-- First Test --")
+    print(f"PSNR value is {psnr(original, contrast)} dB")
     
-	diff = orignalPixelAt - compressedPixelAt
+    # # Value expected: 31.53dB (Wikipedia Example)
-	error = np.sum(np.abs(diff) ** 2)
+    print("\n-- Second Test --")
-
+    print(f"PSNR value is {psnr(original2, contrast2)} dB")
-	error = error/(height*width)
-
-	#MSR = error_sum/(height*width)
-	PSNR = -(10*math.log10(error/(255*255)))
-
-	print("PSNR value is {}".format(PSNR))
 
 
 if __name__ == '__main__':
     main()
-

arithmetic_analysis/bisection.py

@@ -15,7 +15,7 @@ def bisection(function, a, b):  # finds where the function becomes 0 in [a,b] us
         return
     else:
         mid = (start + end) / 2
-        while abs(start - mid) > 0.0000001:  # until we achieve precise equals to 10^-7
+        while abs(start - mid) > 10**-7:  # until we achieve precise equals to 10^-7
             if function(mid) == 0:
                 return mid
             elif function(mid) * function(start) < 0:
@@ -29,5 +29,5 @@ def bisection(function, a, b):  # finds where the function becomes 0 in [a,b] us
 def f(x):
     return math.pow(x, 3) - 2*x - 5
 
-
+if __name__ == "__main__":
     print(bisection(f, 1, 1000))

arithmetic_analysis/intersection.py

@@ -5,12 +5,13 @@ def intersection(function,x0,x1): #function is the f we want to find its root an
     x_n1 = x1
     while True:
         x_n2 = x_n1-(function(x_n1)/((function(x_n1)-function(x_n))/(x_n1-x_n)))
-        if abs(x_n2 - x_n1)<0.00001 :
+        if abs(x_n2 - x_n1) < 10**-5:
             return x_n2
         x_n=x_n1
         x_n1=x_n2
 
 def f(x):
-    return math.pow(x,3)-2*x-5
+    return math.pow(x , 3) - (2 * x) -5
 
+if __name__ == "__main__":
     print(intersection(f,3,3.5))

arithmetic_analysis/lu_decomposition.py

@@ -1,13 +1,14 @@
+# lower–upper (LU) decomposition - https://en.wikipedia.org/wiki/LU_decomposition
 import numpy
 
 def LUDecompose (table):
-    #table that contains our data
+    # Table that contains our data
-    #table has to be a square array so we need to check first
+    # Table has to be a square array so we need to check first
     rows,columns=numpy.shape(table)
     L=numpy.zeros((rows,columns))
     U=numpy.zeros((rows,columns))
     if rows!=columns:
-        return
+        return []
     for i in range (columns):
         for j in range(i-1):
             sum=0
@@ -22,13 +23,10 @@ def LUDecompose (table):
             U[i][j]=table[i][j]-sum1
     return L,U
 
-
+if __name__ == "__main__":
-
+    matrix =numpy.array([[2,-2,1],
-
+                         [0,1,2],
-
+                         [5,3,1]])
-
-
-matrix =numpy.array([[2,-2,1],[0,1,2],[5,3,1]])
     L,U = LUDecompose(matrix)
     print(L)
     print(U)

arithmetic_analysis/newton_method.py

@@ -1,15 +1,18 @@
+# Newton's Method - https://en.wikipedia.org/wiki/Newton%27s_method
+
 def newton(function,function1,startingInt): #function is the f(x) and function1 is the f'(x)
   x_n=startingInt
   while True:
       x_n1=x_n-function(x_n)/function1(x_n)
-      if abs(x_n-x_n1)<0.00001:
+      if abs(x_n-x_n1) < 10**-5:
           return x_n1
       x_n=x_n1
       
 def f(x):
-    return (x**3)-2*x-5
+    return (x**3) - (2 * x) -5
 
 def f1(x):
     return 3 * (x**2) -2
 
+if __name__ == "__main__":
     print(newton(f,f1,3))

data_structures/graph/dijkstra.py

@@ -21,7 +21,7 @@ def minDist(mdist, vset, V):
 def Dijkstra(graph, V, src):
 	mdist=[float('inf') for i in range(V)]
 	vset = [False for i in range(V)]
-	mdist[src] = 0.0;
+	mdist[src] = 0.0
 	
 	for i in range(V-1):
 		u = minDist(mdist, vset, V)