英文:
How do I apply a logical operation between numpy arrays of different sizes?
问题
我有一个存储为形状为（x，y，3）的numpy数组的图像。我有一个形状为（x，y）的numpy ndarray的掩码。如何将该掩码应用于我的图像？cv2.bitwise_and()似乎不起作用，直接使用image[mask]也不行。具体来说，我的掩码包含True和False，我想执行image & mask，但是要一次应用于所有图层而不是逐个图层（R，G，B）处理，然后再将它们合并。
英文:
I have an image stored as a numpy array with the shape (x,y,3). I have a mask that is a numpy ndarray of shape (x,y). How do I apply that mask to my image? cv2.bitwise_and() doesn't seem to work and neither does directly indexing with image[mask]. To be specific, my mask contains True and False and I want to perform image & mask, but applied to all layers at once instead of doing it one layer at a time (R,G,B) and combining them later.
答案1
得分: 1
使用 mask[:.:,None]。这将使 mask 变成一个维度为 (x, y, 1) 的数组，正常的广播规则可以处理 (x, y, 1) 与 (x, y, 3) 的按位与操作。
英文:
Try using mask[:.:,None]. This makes mask out to be an array of dimension (x, y, 1), and the normal broadcasting rules can handle anding (x, y, 1) with (x, y, 3).
通过集体智慧和协作来改善编程学习和解决问题的方式。致力于成为全球开发者共同参与的知识库，让每个人都能够通过互相帮助和分享经验来进步。
评论