High-quality video view interpolation using a layered representation

By C. Lawrence, Zitnick Sing, B. Kang, M. Uyttendaele, Simon A. J. Winder, Richard Szeliski. ACM Transactions on Graphics, 2004

1,665 citations recorded by the source corpus.

Subfields: Computer Vision and Pattern Recognition, Signal Processing

View Paper: https://www.semanticscholar.org/paper/a285a442317f24068f7cc262531976e82a5c41d6

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Vision and Pattern Recognition

In Signal Processing

Citations in the corpus

D-NeRF: Neural Radiance Fields for Dynamic Scenes (2020, background): “For instance, [2, 54, 15] capture the dynamic scene at the same time instant from multiple views, to then generate 4D space-time visualizations.”
Ieee Transactions on Pattern Analysis and Machine Intelligence 1 Stereo Processing by Semi-global Matching and Mutual Information (methodology): “Almost all of the currently top-ranked algorithms [2], [3], [7], [9], [11]–[15] on the Tsukuba, Venus, Teddy and Cones data set [18] optimize a global energy function.”
NeX: Real-time View Synthesis with Neural Basis Expansion (2021, background): “[24], which combines depth estimation with soft blending of an estimated geometry, and other 3D reconstruction based techniques [54, 10].”
High performance imaging using large camera arrays (2005, methodology): “For example, segmentation-based stereo methods have recently been proven very useful for spatial view interpolation [Zitnick et al. 2004] and analysis of structure and motion in dynamic scenes [Tao et al.”
DeepView: View Synthesis With Learned Gradient Descent (2019, methodology): “Other approaches explicitly reconstruct 3D geometry, such as a global 3D reconstruction [15] or a collection of perinput-view depth maps [37, 6, 20] using multi-view stereo algorithms.”
Video Frame Interpolation via Adaptive Separable Convolution (2017, methodology): “Generic image-based rendering algorithms can also be used to improve frame synthesis results [33, 66].”
Stereo magnification (2018, background): “Many approaches focus on the interpolation setting, and operate by either interpolating rays from dense imagery (“light field rendering”) [Gortler et al. 1996; Levoy and Hanrahan 1996], or reconstructing scene geometry from sparse views [Debevec et al. 1996; Hedman et al. 2017; Zitnick et al. 2004].”
Unsupervised Learning of Depth and Ego-Motion from Video (2017, background): “A classic paradigm for view synthesis is to ﬁrst either estimate the underlying 3D geometry explicitly or establish pixel correspondence among input views, and then synthesize the novel views by compositing image patches from the input views (e.g., [4, 55, 43, 6, 9]).”
BARF: Bundle-Adjusting Neural Radiance Fields (2021, background): “The task has been closely tied to 3D reconstruction since its introduction [7, 72, 18].”
Image Inpainting : Overview and Recent Advances (2014, background): “Figure 12 illustrates this application by showing in (a) a reference view (view four of the multiview sequence called ballet (see [67]) and in (c) the result of the projection on a virtual viewpoint corresponding to a displacement to the right of the camera.”
Deep blending for free-viewpoint image-based rendering (2018, background): “The importance of maintaining accurate results at occlusion edges was identiied early, using soft visibility [Pulli et al. 1997; Zheng et al. 2009] or superpixels and alpha matting [Zitnick et al. 2004] to improve the blending.”
Learning Conditional Random Fields for Stereo (2007, background): “Taken to an extreme, this translates into the assumption that disparity jumps always coincide with color edges, which is the basis of a large number of recent segmentation-based stereo methods [10, 14–19].”
3-D Video Representation Using Depth Maps (2011, methodology): “In advanced synthesis methods, a reliability-based approach is taken with one [ 52 ] or two [33] boundary layers.”
Single-View View Synthesis With Multiplane Images (2020, methodology): “A number of classical approaches to this problem have been explored [10, 16], including methods that involve estimation of local geometric proxies [1, 11, 23, 38].”
Nonlinear disparity mapping for stereoscopic 3D (2010, methodology): “The classical approach to this problem has been to perform imagebased view interpolation, which either requires a very large number of densely sampled input images or additional accurate depth maps to achieve high quality results [Gortler et al. 1996; Levoy and Hanrahan 1996; Shade et al. 1998; Zitn”
Neural Body: Implicit Neural Representations with Structured Latent Codes for Novel View Synthesis of Dynamic Humans (2020, background): “To extend the range, [62, 6, 45] infer depth maps from input images as proxy geometries.”