machine learning - Is there a layer in Caffe that can get an arbitrary sub block in a blob? -
is there layer in caffe can arbitrary sub block in blob?
afaik there not general slicing layer in caffe.
if want extract sub-blocks specific channels of blob, can use "slice"
layer.
depending on requirements, might find "crop"
layer sufficient needs.
if need more flexible access sub-blocks, might want consider using "python"
layer.
an example using "crop"
layer
pointed out "crop"
layer expects 2 "bottom"s since second 1 used reference shape, can produce using "dummydata"
layer.
suppose want select x[:,:,3:20,5:40]
, 17x35 crop
layer { name: "ref_shape_17x35" type: "dummydata" top: "ref_shape_17x35" dummy_data_param { shape { dim: 1 dim: 1 dim: 17 dim: 35 } } } layer { name: "crop_x" type: "crop" bottom: "x" bottom: "ref_shape_17x35" top: "crop_x" crop_param { axis: 2 # not crop first 2 dimensions offset: 3 offset: 5 } }
i did not try myself, should work (let me know if doesn't).
Comments
Post a Comment