如何找到矩阵列表的置换分布以计算图像的香农熵

Question

其中 H(X)=−∑xp(x)logp(x) 是 Shanon entropy.In 论文“History of art paintings through the lens of entropy and complexity”

Answer 1

我最近在研究绘画数据集的探索性工作，希望这能解决您的问题。根据参考文献分布长度不能超过24.

def permutataion_distribution(matrix_list):
'''
Returns the distribution of matrix permutaions.
Input: 
    matrix_list: the output of sliding_matrix.
output: 
    num_of_perm: list of number of each permutations in order recieved
    list_of_perm: list of permutations in order recieved
    distribution: distrubution
    
'''
list_of_perm = []
num_of_perm = []
distribution = np.zeros(24)
for i in matrix_list:
    mat_perm = np.argsort((np.array[i]).flatten())
    indicator = 0
    for j in range(0, len(list_of_perm)):
        if np.all(mat_perm == list_of_perm[j]):
            num_of_perm[j] = num_of_perm[j] + 1
            indicator = 1
    if indicator == 0:
        list_of_perm.append(mat_perm)
        num_of_perm.append(1)
        
distribution = np.array(num_of_perm) / sum(num_of_perm)
return num_of_perm, list_of_perm, distribution

如何找到矩阵列表的置换分布以计算图像的香农熵

How to find the permutation distribution for a list of matrices for calculating Shanon entropy of an Image

python

machine-learning