优化python中函数的性能

我需要根据B从A中提取一个值，但是使用for循环对于大型数据集来说是不切实际的.虽然下面的代码避免使用for循环，但它仍然需要相同的时间.

import numpy as np

# Given matrices A and B
A = np.array([[254, 0, 0],
              [109, 0, 1],
              [126, 0, 2],
              [66, 0, 3],
              [220, 1, 0],
              [98, 1, 1],
              [230, 1, 2],
              [17, 1, 3],
              [83, 2, 0],
              [106, 2, 1],
              [123, 2, 2],
              [57, 2, 3]])

B = np.array([[1, 2],
              [0, 1],
              [1, 0],
              [1, 1],
              [-1, 2],
              [1, 3],
              [1, 1],
              [0, 0],
              [2, 2],
              [1, 0],
              [2, 3],
              [0, 1]])

## these two functions give the same results.
def get_pixel2d(A, B): 
    corresponding_rows = np.all(A[:, 1:3] == B[:, None], axis=-1)
    get_pixel_final = (corresponding_rows * A[:, 0]).sum(axis=1)
    return get_pixel_final

# this is more faster
def get_pixel2d(A, B):
    corresponding_rows = np.all(A[:, 1:3] == B[:, None], axis=-1)
    get_pixel_final = np.sum(A[:, 0] * corresponding_rows, axis=1)
    return get_pixel_final

result = get_pixel2d(A, B)
print(result)

[230 109 220  98   0  17  98 254 123 220  57 109]