Question

我只是想知道这是否是一种加快Python中for循环性能的方法。

for i in range (0,img.shape[0],new_height):
    for j in range(0,img.shape[1],new_width):
        cropped_image = img[i:i+new_height,j:j+new_width]
        yuv_image = cv2.cvtColor(cropped_image,cv2.COLOR_BGR2YUV)
        Y,U,V = cv2.split(yuv_image)
        pixel_image_y = np.array(Y).flatten()

Answer 1

将整个图像转换为YUV空间后，我们可以简单地重塑成较小的块-

m,n = img.shape[:2]
yuv = cv2.cvtColor(img,cv2.COLOR_BGR2YUV)
yuv4D = yuv[...,0].reshape(m//new_height,new_height,n//new_width,new_width)
out = yuv4D.swapaxes(1,2).reshape(-1,new_height*new_width)

在1024x1024 RGB图像上的时间-

In [157]: img = np.random.randint(0,256,(1024,1024,3)).astype(np.uint8)
     ...: new_height,new_width = 32,32

In [158]: %%timeit
     ...: out = []
     ...: for i in range (0,img.shape[0],new_height):
     ...:     for j in range(0,img.shape[1],new_width):
     ...:         cropped_image = img[i:i+new_height,j:j+new_width]
     ...:         yuv_image = cv2.cvtColor(cropped_image,cv2.COLOR_BGR2YUV)
     ...:         Y,U,V = cv2.split(yuv_image)
     ...:         pixel_image_y = np.array(Y).flatten()
     ...:         out.append(pixel_image_y)
11.9 ms ± 991 µs per loop (mean ± std. dev. of 7 runs, 100 loops each)

In [159]: %%timeit
     ...: m,n = img.shape[:2]
     ...: yuv = cv2.cvtColor(img,cv2.COLOR_BGR2YUV)
     ...: yuv4D = yuv[...,0].reshape(m//new_height,new_height,n//new_width,new_width)
     ...: out1 = yuv4D.swapaxes(1,2).reshape(-1,new_height*new_width)
1.48 ms ± 5.23 µs per loop (mean ± std. dev. of 7 runs, 1000 loops each)

有没有一种方法可以加快python中的嵌套for循环？

1 个答案: