我有一个mask
数组,它代表一个二维二进制图像。假设它很简单:
mask = np.zeros((9, 9), dtype=np.uint8)
# 0 0 0 | 0 0 0 | 0 0 0
# 0 0 0 | 0 0 0 | 0 0 0
# 0 0 0 | 0 0 0 | 0 0 0
# ------+-------+------
# 0 0 0 | 0 0 0 | 0 0 0
# 0 0 0 | 0 0 0 | 0 0 0
# 0 0 0 | 0 0 0 | 0 0 0
# ------+-------+------
# 0 0 0 | 0 0 0 | 0 0 0
# 0 0 0 | 0 0 0 | 0 0 0
# 0 0 0 | 0 0 0 | 0 0 0
假设我想翻转ninth左中的元素:
# 0 0 0 | 0 0 0 | 0 0 0
# 0 0 0 | 0 0 0 | 0 0 0
# 0 0 0 | 0 0 0 | 0 0 0
# ------+-------+------
# 1 1 1 | 0 0 0 | 0 0 0
# 1 1 1 | 0 0 0 | 0 0 0
# 1 1 1 | 0 0 0 | 0 0 0
# ------+-------+------
# 0 0 0 | 0 0 0 | 0 0 0
# 0 0 0 | 0 0 0 | 0 0 0
# 0 0 0 | 0 0 0 | 0 0 0
我不正确的方法是这样的:
x = np.arange(mask.shape[0])
y = np.arange(mask.shape[1])
mask[np.logical_and(y >= 3, y < 6), x < 3] = 1
# 0 0 0 | 0 0 0 | 0 0 0
# 0 0 0 | 0 0 0 | 0 0 0
# 0 0 0 | 0 0 0 | 0 0 0
# ------+-------+------
# 1 0 0 | 0 0 0 | 0 0 0
# 0 1 0 | 0 0 0 | 0 0 0
# 0 0 1 | 0 0 0 | 0 0 0
# ------+-------+------
# 0 0 0 | 0 0 0 | 0 0 0
# 0 0 0 | 0 0 0 | 0 0 0
# 0 0 0 | 0 0 0 | 0 0 0
(这是我真正要处理的约束的简化,在这种情况下,它很难表示为mask[:3,3:6] = 1
之类的东西。请考虑约束是任意的,例如x % 2 == 0 && y % 3 == 0
会。
当两个索引数组的形状相同时,Numpy的行为是将它们成对放置,最终只能选择上面的3个元素,而不是我想要的9个。
如何使用适用于不同轴的约束条件更新正确的元素?鉴于约束是独立的,是否可以仅通过评估我的约束N + M次而不是N * M来做到这一点?
答案 0 :(得分:1)
根据您的评论,让我们尝试以下更出色的示例:
mask = np.zeros((90,90), dtype=np.uint8)
# criteria
def f(x,y): return ((x-20)**2 < 50) & ((y-20)**2 < 50)
# ranges
x,y = np.arange(90), np.arange(90)
# meshgrid
xx,yy = np.meshgrid(x,y)
zz = f(xx,yy)
# mask
mask[zz] = 1
plt.imshow(mask, cnap='gray')
输出:
答案 1 :(得分:1)
您无法broadcast
布尔数组,但可以使用ix_
构造等效的数字索引:
In [330]: np.ix_((y>=3)&(y<6), x<3)
Out[330]:
(array([[3],
[4],
[5]]), array([[0, 1, 2]]))
应用:
In [331]: arr = np.zeros((9,9),int)
In [332]: arr[_330] = 1
In [333]: arr
Out[333]:
array([[0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0],
[1, 1, 1, 0, 0, 0, 0, 0, 0],
[1, 1, 1, 0, 0, 0, 0, 0, 0],
[1, 1, 1, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0]])
尝试broadcast
的布尔值会直接引发错误(too many indices
):
arr[((y>=3)&(y<6))[:,None], x<3]