如何找到两幅图像之间的相关性。

huangapple
go评论109阅读模式
英文:

How to find correlation between two images

问题

I need to find correlation between two images, using numpy, but basic math only. I have the problem: "IndexError: index 5434 is out of bounds for axis 0 with size 5434". And i have a code. Tell me what to do, pls.

  1. img = PIL.Image.open("SR1.png").convert("L")
  2. im = numpy.array(img)
  3. img2 = PIL.Image.open("SR61.png").convert("L")
  4. im2 = numpy.array(img2)
  5. np.array(im,dtype=float)
  6. np.array(im2,dtype=float)
  7. import math
  8. import cmath
  9. def correlationCoefficient(X, Y, n) : 
  10.     sum_X = 0
  11.     sum_Y = 0
  12.     sum_XY = 0
  13.     squareSum_X = 0
  14.     squareSum_Y = 0
  17.     i = 0
  18.     for i in range(n) : 
  19.         sum_X = sum_X + X[i]
  20.         sum_Y = sum_Y + Y[i] 
  21.         sum_XY = sum_XY + X[i] * Y[i] 
  22.         squareSum_X = squareSum_X + X[i] * X[i] 
  23.         squareSum_Y = squareSum_Y + Y[i] * Y[i] 
  25.         i = i + 1
  27.     corr = (n * sum_XY - sum_X * sum_Y)/(cmath.sqrt((n * squareSum_X - sum_X * sum_X)* (n * squareSum_Y - sum_Y * sum_Y))) 
  28.     return corr
  31. X = im.flatten()
  32. Y = im2.flatten()
  35. n = len(X) 
  38. print ('{0:.6f}'.format(correlationCoefficient(X, Y, n)))

Please note that I've provided the code without the translation parts as requested. If you have any specific questions or need further assistance, please let me know.

英文:

I need to find correlation between two images, using numpy, but basic math only. I have the problem:"***
IndexError: index 5434 is out of bounds for axis 0 with size 5434***". And i have a code. Tell me what to do, pls.

  1. img = PIL.Image.open("SR1.png").convert("L")
  2. im = numpy.array(img)
  3. img2 = PIL.Image.open("SR61.png").convert("L")
  4. im2 = numpy.array(img2)
  5. np.array(im,dtype=float)
  6. np.array(im2,dtype=float)
  7. import math
  8. import cmath
  9. def correlationCoefficient(X, Y, n) : 
  10.     sum_X = 0
  11.     sum_Y = 0
  12.     sum_XY = 0
  13.     squareSum_X = 0
  14.     squareSum_Y = 0
  15.   
  16.   
  17.     i = 0
  18.     for i in range(n) : 
  19.         sum_X = sum_X + X[i]
  20.         sum_Y = sum_Y + Y[i] 
  21.         sum_XY = sum_XY + X[i] * Y[i] 
  22.         squareSum_X = squareSum_X + X[i] * X[i] 
  23.         squareSum_Y = squareSum_Y + Y[i] * Y[i] 
  24.       
  25.         i = i + 1
  27.     corr = (n * sum_XY - sum_X * sum_Y)/(cmath.sqrt((n * squareSum_X - sum_X * sum_X)* (n * squareSum_Y - sum_Y * sum_Y))) 
  28.     return corr
  29.  
  30.   
  31. X = im.flatten()
  32. Y = im2.flatten()
  33.    
  35. n = len(X) 
  38. print ('{0:.6f}'.format(correlationCoefficient(X, Y, n)))

如何找到两幅图像之间的相关性。

答案1

得分: 2

你可以使用numpy中的corrcoef函数来计算皮尔逊相关系数。首先需要将两个图像数组都使用flatten函数展平：

  1. np.corrcoef(im1.flatten(), im2.flatten())
英文:

You can use the function corrcoef in numpy to find Peason correlation. First you need to flatten both image arrays:

  1. np.corrcoef(im1.flatten(), im2.flatten())

答案2

得分: 2

以下是您功能的矢量化版本：

  1. import numpy as np
  3. def correlationCoefficient(X, Y):
  4.     n = X.size
  5.     sum_X = X.sum()
  6.     sum_Y = Y.sum()
  7.     sum_XY = (X*Y).sum()
  8.     squareSum_X = (X*X).sum()
  9.     squareSum_Y = (Y*Y).sum()
  10.     corr = (n * sum_XY - sum_X * sum_Y)/(np.sqrt((n * squareSum_X - sum_X * sum_X)* (n * squareSum_Y - sum_Y * sum_Y))) 
  11.     return corr

还重要的是要对图像数组进行归一化以避免溢出：

  1. from PIL import Image
  3. img1 = Image.open("1.jpg").convert("L")
  4. im1 = np.array(img1)/255
  6. img2 = Image.open("2.jpg").convert("L")
  7. im2 = np.array(img2)/255
  9. print ('{0:.6f}'.format(correlationCoefficient(im1, im2)))
英文:

Here's a vectorized version of your function:

  1. import numpy as np
  3. def correlationCoefficient(X, Y):
  4.     n = X.size
  5.     sum_X = X.sum()
  6.     sum_Y = Y.sum()
  7.     sum_XY = (X*Y).sum()
  8.     squareSum_X = (X*X).sum()
  9.     squareSum_Y = (Y*Y).sum()
  10.     corr = (n * sum_XY - sum_X * sum_Y)/(np.sqrt((n * squareSum_X - sum_X * sum_X)* (n * squareSum_Y - sum_Y * sum_Y))) 
  11.     return corr

It is also important to normalize your image arrays to avoid overflow:

  1. from PIL import Image
  3. img1 = Image.open("1.jpg").convert("L")
  4. im1 = np.array(img1)/255
  6. img2 = Image.open("2.jpg").convert("L")
  7. im2 = np.array(img2)/255
  9. print ('{0:.6f}'.format(correlationCoefficient(im1, im2)))

huangapple
  • 本文由 发表于 2020年1月6日 15:53:13
  • 转载请务必保留本文链接：https://go.coder-hub.com/59608470.html
  • computer-vision
  • machine-learning
  • numpy
  • python
  • python-3.x
匿名

发表评论

匿名网友

:?: :razz: :sad: :evil: :!: :smile: :oops: :grin: :eek: :shock: :???: :cool: :lol: :mad: :twisted: :roll: :wink: :idea: :arrow: :neutral: :cry: :mrgreen:

确定