reg23.autograd Namespace Reference

Functions
	project_drr_setup_context (ctx, inputs, output)
	project_drr_backward (ctx, d_loss_d_drr)
	normalised_cross_correlation_setup_context (ctx, inputs, output)
	normalised_cross_correlation_backward (ctx, d_loss_d_zncc, *args)

Variables
	project_drr_backward
	setup_context

Function Documentation

normalised_cross_correlation_backward()

reg23.autograd.normalised_cross_correlation_backward	(		ctx,
			d_loss_d_zncc,
		*	args
	)

:param ctx:
:param d_loss_d_zncc: tensor of size (,); the derivative of the loss w.r.t. the ZNCC
:return: tensor of the same size as a: the derivative of the loss w.r.t. the tensor `a`

normalised_cross_correlation_setup_context()

reg23.autograd.normalised_cross_correlation_setup_context	(		ctx,
			inputs,
			output
	)

project_drr_backward()

reg23.autograd.project_drr_backward	(		ctx,
			d_loss_d_drr
	)

:param ctx:
:param d_loss_d_drr: tensor of size (drr_height, drr_width); the derivative of the loss w.r.t. the DRR
:return: tensor of size (4, 4): the derivative of the loss w.r.t. the inverse homography matrix

project_drr_setup_context()

reg23.autograd.project_drr_setup_context	(		ctx,
			inputs,
			output
	)

Variable Documentation

project_drr_backward

reg23.autograd.project_drr_backward

setup_context

reg23.autograd.setup_context