Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Image Jacobians #351

Closed
BenedictChannn opened this issue Jul 11, 2023 · 1 comment
Closed

Image Jacobians #351

BenedictChannn opened this issue Jul 11, 2023 · 1 comment

Comments

@BenedictChannn
Copy link

Hi all,

If my residual is a pixel of a CV image (indexed sf.DataBuffer), the Jacobians and Hessians are set to zero. I need to account for the gradient of the CV image and also the Jacobian of the reprojection of a 3D point from a reference frame to the current frame. Defining a custom Jacobian would not be ideal since I am dealing with a lot of symbolic variables. Is there a way to deal with this? Especially since the CV image is now represented as a 1D array instead the standard Matrix.

image
Ideally, the autogenerated Jacobians should include the image intensity gradient and the Jacobian of the warping function as well.

Also, is there a way to handle conditional statements for symbolic computations? Are there other similar functions/ capabilities like those in logic.py

I also noticed that when using sf.logical_or operations with more than 2 arguments (unsafe = True) ,output is max(True, False) when it should by right be True?

I would appreciate any help on the above!! Thanks.

@aaron-skydio
Copy link
Member

Addressed here, will continue discussion there: #348 (comment)

In particular, this is not a bug - a DataBuffer is a piecewise constant function of its index, and if you want a notion of an image gradient you need to represent that

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants