Space-Time Completion of Video

By Yonatan Wexler, Eli Shechtman, Michal Irani. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2007

758 citations recorded by the source corpus.

Subfields: Computer Vision and Pattern Recognition

View Paper: https://www.semanticscholar.org/paper/3ace907ea90aed1fb4287b78b19a9f65a462d247

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

Citations in the corpus

Flow-edge Guided Video Completion (2020, methodology): “Up until a few years ago, most methods used patch-based synthesis techniques [14,26,39].”
MSG-GAN: Multi-Scale Gradients for Generative Adversarial Networks (2019, methodology): “Multi-scale image generation is a well established technique, with methods existing well before deep networks became popular for this task [20, 35].”
MSG-GAN: Multi-Scale Gradient GAN for Stable Image Synthesis (2019, methodology): “The idea of multi-scale image generation is a well established technique, with methods existing well before deep networks became popular for this task [17, 29].”
Video-to-Video Synthesis (2018, methodology): “Our method is inspired by previous application-speciﬁc video synthesis methods [58,60,61,75].”
Transformation-Grounded Image Generation Network for Novel 3D View Synthesis (2017, background): “Traditional image completion or hole filling methods often exploit local image information [9, 2, 45] and have shown impressive results for filling small holes or texture synthesis.”
Globally and locally consistent image completion (2017, methodology): “Although many approaches have been proposed for image completion, such as patch-based image synthesis [Barnes et al. 2009; Darabi et al. 2012; Huang et al. 2014; Simakov et al. 2008; Wexler et al. 2007], it remains a challenging problem because it often requires high-level recognition of scenes.”
3D Photography Using Context-Aware Layered Depth Inpainting (2020, methodology): “…synthesis [9, 8], example-based methods complete the missing regions by transferring the contents from the known regions of the image, either through non-parametric patch-based synthesis [63, 1, 5, 20] or solving a Markov Random Field model using belief propagation [26] or graph cut [48, 28, 16].”
Image melding (2012, methodology): “Therefore previous methods [Wexler et al. 2007; Simakov et al. 2008; Barnes et al. 2009] used smaller patches (e.g., 5 × 5 or 7 × 7) that generally lead to sharper and more flexible synthesis (linear structures can slightly bend to better connect) and the expense structural changes.”
Summarizing visual data using bidirectional similarity (2008, methodology): “(2), our similarity measure reduces to the Coherence objective function of [18], and our optimization algorithm of Sec.”
Transformation guided image completion (2013, methodology): “, coordinate descent) [25, 15] or MRF models [14, 13] which can be solved efficiently using belief propagation.”
Video Inpainting of Complex Scenes (2014, methodology): “This phenomenon was also noted in [40].”
Hierarchical Style-based Networks for Motion Synthesis (2020, background): “] to facilitate this task. Our work is dierent from this branch of work because there exists large motion gap between start and end states in our settings. Another branch of work is video completion [4,19,32]. It receives two nonconsecutive frames as input and aims to ll the motion gap between start”