A dual-branch optical flow network using a 90-degree rotated 'orthogonal' view reduces polar distortion errors and sets new state-of-the-art results on MPFDataset and FlowScape.
Omniflownet: a perspective neural network adaptation for optical flow estimation in om- nidirectional images
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
PriOr-Flow: Enhancing Primitive Panoramic Optical Flow with Orthogonal View
A dual-branch optical flow network using a 90-degree rotated 'orthogonal' view reduces polar distortion errors and sets new state-of-the-art results on MPFDataset and FlowScape.