RGB 像素的熵

问题描述

如何在 Python 中正确地获取单个 RGB 像素的熵，或如下图所示。

试过这个：

def entropy(*X):
    return  np.sum(-p * np.log2(p) if p > 0 else 0 for p in
        (np.mean(functools.reduce(np.logical_and,(predictions == c for predictions,c in zip(X,classes))))
            for classes in itertools.product(*[set(x) for x in X])))

print('Entropy:',entropy(np.array([255,2,2]),np.array([255,1]),0]),1,0])))
print('Entropy:',2])))
print('Entropy:',entropy(np.array([0,255,100,200])))

结果：

Entropy: 1.584962500721156
Entropy: 0.9182958340544896
Entropy: 0.9182958340544896
Entropy: 0.9182958340544896
Entropy: 1.584962500721156

但是红色和绿色像素的熵相同，似乎无法区分，并且看起来与下表不同。我做错了什么以及如何正确地做到这一点？

来源：https://www.researchgate.net/publication/282687781_A_methodology_for_using_crowdsourced_data_to_measure_uncertainty_in_natural_speech

解决方法

如果有帮助，试试这个

def entropy(labels,base=None):
    value,counts = np.unique(labels,return_counts=True)
    norm_counts = counts / counts.sum()
    print(norm_counts)
    e = 2.718281828
    base = e if base is None else base
    return -(norm_counts * np.log(norm_counts)/np.log(base)).sum()

f = np.array([[[[255,10,2],[255,1],0],1,2,0]]]])
h = entropy(f,2)
print('h=',h)