We can use the following example(ROIPooling exmple from MxNet) to test:
import mxnet as mx
x = mx.nd.array([[[[ 0., 1., 2., 3., 4., 5.],
[ 6., 7., 8., 9., 10., 11.],
[ 12., 13., 14., 15., 16., 17.],
[ 18., 19., 20., 21., 22., 23.],
[ 24., 25., 26., 27., 28., 29.],
[ 30., 31., 32., 33., 34., 35.],
[ 36., 37., 38., 39., 40., 41.],
[ 42., 43., 44., 45., 46., 47.]]]])
// region of interest i.e. bounding box coordinates.
y = mx.nd.array([[0,0,0,3,3]])
// returns array of shape (2,2) according to the given roi with max pooling.
mx.nd.ROIPooling(x, y, (2,2), 1.0) = [[[[ 7., 9.],
[ 19., 21.]]]]
mx.nd.ROIAlign(x, y, (2, 2), 1.0) = [[[[5.25, 6.75],
[14.25, 15.75]]]]
I think the result of mx.nd.ROIAlign(x, y, (2, 2), 1.0) should be the same as
mx.nd.ROIPooling(x, y, (2,2), 1.0).
[ Full content available at:
https://github.com/apache/incubator-mxnet/issues/12403 ]
This message was relayed via gitbox.apache.org for [email protected]
