Full explicit consistency constraints in uncalibrated multiple homography estimation
This work addresses a fundamental issue in computer vision for applications like 3D reconstruction and image stitching, though it appears incremental as it builds on existing homography estimation methods.
The paper tackled the problem of ensuring that a set of 3x3 matrices represent genuine homographies for multiple planes between two views by revealing a complete set of constraints, and it showed how to use these constraints to obtain more accurate homography estimates, resolving a long-standing research question.
We reveal a complete set of constraints that need to be imposed on a set of 3-by-3 matrices to ensure that the matrices represent genuine homographies associated with multiple planes between two views. We also show how to exploit the constraints to obtain more accurate estimates of homography matrices between two views. Our study resolves a long-standing research question and provides a fresh perspective and a more in-depth understanding of the multiple homography estimation task.