Understanding Deconvolution Layer and Crop Layer.

[William Tan]

Hi all,

I have a few questions about Deconvolution Layer and Crop Layer.

1. How do I calculate the output size w.r.t. input size, filter size, padding, and strides?

2. What is the purpose of Crop Layer?

Regards,

William

[William Tan]

Sorry, the first question is referring to the Deconvolution Layer.

[Evan Shelhamer]

1. How do I calculate the output size w.r.t. input size, filter size, padding, and strides?

​The exact logic is handled by `compute_output_shape()`: https://github.com/BVLC/caffe/blob/master/src/caffe/layers/deconv_layer.cpp#L7-L22​

2. What is the purpose of Crop Layer?

It adjusts the coordinates betweens layers to align their outputs for fusion or mapping output pixels to their corresponding input pixels.

See PRs https://github.com/BVLC/caffe/pull/3613 and https://github.com/BVLC/caffe/pull/3570 for more detail and background.

Evan Shelhamer

--
You received this message because you are subscribed to the Google Groups "Caffe Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to caffe-users...@googlegroups.com.
To post to this group, send email to caffe...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/caffe-users/cd15d6b7-a087-4f99-9af9-c20fddc2980a%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.