DSER combines spectral epipolar regularization with a hybrid pipeline of gradient initialization, plane-sweeping, multiscale refinement, and occlusion-aware random walk to produce structurally consistent depth maps from light fields.
Unsupervised depth estimation of light fields with 3D convolutional neural networks.IEEE Trans
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2025 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
DSER: Spectral Epipolar Representation for Efficient Light Field Depth Estimation
DSER combines spectral epipolar regularization with a hybrid pipeline of gradient initialization, plane-sweeping, multiscale refinement, and occlusion-aware random walk to produce structurally consistent depth maps from light fields.