Open FrederikWarburg opened 2 years ago
Hi
As I understand, eq. 6 in arxiv paper should correspond to line: https://github.com/XinJCheng/CSPN/blob/b3e487bdcdcd8a63333656e69b3268698e543181/cspn_pytorch/models/cspn.py#L81 However, raw_depth_input = blur_depth which is the predicted depth by the network. I wondering if
if sparse_depth is not None: result_depth = ( 1 - sparse_mask ) * result_depth + sparse_mask * raw_depth_input
should be
if sparse_depth is not None: result_depth = ( 1 - sparse_mask ) * result_depth + sparse_mask * sparse_depth
such that you keep the depth at the points where you have depth?
Hi
As I understand, eq. 6 in arxiv paper should correspond to line: https://github.com/XinJCheng/CSPN/blob/b3e487bdcdcd8a63333656e69b3268698e543181/cspn_pytorch/models/cspn.py#L81 However, raw_depth_input = blur_depth which is the predicted depth by the network. I wondering if
should be
such that you keep the depth at the points where you have depth?