Open andrea-unity opened 2 years ago
With some modifications it can support right->left disparity. The correlation sampler returned by corr_block()
assumes left->right disparity so the order of the input feature maps will need to be flipped. You should re-train the model since L->R disparity is negative and R->L disparity is positive.
Thank you. I will have a look at the code and I will come back to you.
The reason that I would like the left and the right disparity is because I want to try to use RAFT-3D but it seems I need to pass as input images + disparity (L+R).
Andrea
On Wed, May 4, 2022 at 8:47 PM Lahav Lipson @.***> wrote:
With some modifications it can support right->left disparity. The correlation sampler returned by corr_block() assumes left->right disparity so the order of the input feature maps will need to be flipped. You should need to re-train the model since L->R disparity is negative and R->L disparity is positive.
— Reply to this email directly, view it on GitHub https://github.com/princeton-vl/RAFT-Stereo/issues/50#issuecomment-1118136501, or unsubscribe https://github.com/notifications/unsubscribe-auth/AYWM66JINN4UKNTZQ5224FLVINAG3ANCNFSM5VDQDX7A . You are receiving this because you authored the thread.Message ID: @.***>
Using your code I'm able to compute the disparity Left, I'm wondering if your algorithm support also up or down disparity? Thank you
Hello
I have the same question, could you please give us more hints how to realize the right-left disparity please? This is also linked to the open-issue 69.
Thank you very much in advance. Hai
@HaiLiExp You can also flip both images horizontally and reverse the input order to get the right-left disparity.
Using your code I'm able to compute the disparity Left, I'm wondering if your algorithm support also Right disparity? Thank you