Adaptive Tiling for Least-Squares Phase Unwrapping: Runtime and Accuracy

The paper studies adaptive tiling strategies for least-squares phase unwrapping, comparing regular grids, quadtree, and kd-tree partitions. It evaluates nine criteria for quadtree subdivision, such as residue count and fringe density, across various tile sizes and budgets. Experiments show that while adaptive partitions use fewer tiles, they are slower than optimized grids and can reduce accuracy, indicating that partition construction and larger tile solving outweigh boundary savings.

Machine-generated by The Flow from the publisher's headline and feed description — not written or checked by a human. The full article lives at arXiv Computer Vision.