How to do minimum pooling in tensorflow?

user_3259896 0
Viewed 1059

Tensorflow has operations for average and max pooling, but not for minimum pooling.

Is there some sort of work around to get min pooling?

keras tensorflow deep-learning machine-learning python
1 Answers

we can manually recreate MaxPooling in this way...

x = np.random.uniform(0,1, (5,30,30,3)).astype('float32')

n_channel = x.shape[-1]
patches = tf.image.extract_patches(x, 
                                   sizes = [1, 3, 3, 1], 
                                   strides = 4*[1], 
                                   rates = 4*[1], 
                                   padding = 'VALID')

channel_pool = [tf.reduce_max(patches[:,:,:,c::n_channel], keepdims=True, axis=-1) for c in range(n_channel)]
res = tf.concat(channel_pool, axis=-1)

tf.reduce_all(res == MaxPool2D(pool_size=(3, 3), strides=(1,1), padding="valid")(x)) ## TRUE !!!

given the example above, we can switch to MinPooling simply using tf.reduce_min

def min_pool(x):

    n_channel = x.shape[-1]
    patches = tf.image.extract_patches(x, 
                                       sizes = [1, 3, 3, 1], 
                                       strides = 4*[1], 
                                       rates = 4*[1], 
                                       padding = 'VALID')

    channel_pool = [tf.reduce_min(patches[:,:,:,c::n_channel], keepdims=True, axis=-1) for c in range(n_channel)]

    return tf.concat(channel_pool, axis=-1)

we can wrap the whole procedure inside a Lambda layer in order to use it inside a keras model: Lambda(min_pool)(x)

user_103750490
Related