glastonburyc/coord-conv-pytorch

All convolution operators replaced with coordinate convolutions

coord-conv-pytorch's Issues

Calculation in radius_channel

Thank you for your sharing. But why this stpe is different? [https://github.com/uber-research/CoordConv/blob/9787d86015d2d32106dc70d2015c2ccb66c662c2/CoordConv.py#L80]

When the coordinates of x, y is normalized to [-1,1], the center of image is [ 0, 0], why you minus 0.5 ? Thanks~

This implement does not support x and y with different dimension.

above

Object Detection Example

Is there an example code integrating this with FasterRCNN as mentioned in the paper?

It seems to me like this solution has potential impact on localization tasks applied to OCR. I'd be curious to see what impact it would have on the ICDAR 2013 Table Competition dataset.

See https://arxiv.org/pdf/1804.06236.pdf for