AppProp: all-pairs appearance-space edit propagation

By Xiaobo An, Fabio Pellacini. 2008

251 citations recorded by the source corpus.

We present an intuitive and efficient method for editing the appearance of complex spatially-varying datasets, such as images and measured materials. In our framework, users specify rough adjustments that are refined interactively by enforcing the policy that similar edits are applied to spatially-close regions of similar appearance. Rather than proposing a specific user interface, our method allows artists to quickly and imprecisely specify the initial edits with any method or workflow they feel most comfortable with. An energy optimization formulation is used to propagate the initial rough adjustments to the final refined ones by enforcing the editing policy over all pairs of points in the dataset. We show that this formulation is equivalent to solving a large linear system defined by a dense matrix. We derive an approximate algorithm to compute such a solution interactively by taking advantage of the inherent structure of the matrix. We demonstrate our approach by editing images, HDR radiance maps, and measured materials. Finally, we show that our framework generalizes prior methods while providing significant improvements in generality, robustness and efficiency.

Subfields: Computational Mechanics, Computer Graphics and Computer-Aided Design, Computer Vision and Pattern Recognition

How researchers see this paper

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

In Computational Mechanics

In Computer Graphics and Computer-Aided Design

In Computer Vision and Pattern Recognition

Citations in the corpus

Rewriting a Deep Generative Model (2020, background): “Edit transfer methods propagate pixel edits to corresponding locations in other images of the same object or adjacent frames in the same video [2,77,28,14,13,81,21].”
Real-time user-guided image colorization with learned deep priors (2017, methodology): “To reduce user efforts, later methods focused on designing better similarity metrics (Luan et al. 2007; Qu et al. 2006) and utilizing long-range connections (An and Pellacini 2008; Xu et al. 2009).”
Transfusive image manipulation (2012, background): “2008], or interactions between distant pixels [An and Pellacini 2008] (see the surveys in e.”
Efficient affinity-based edit propagation using K-D tree (2009, methodology): “While the exact formulations of propagation minimization vary in literature [Lischinski et al. 2006; Pellacini and Lawrence 2007; An and Pellacini 2008], the minimization is typically performed on a per-pixel basis, which makes the computational cost prohibitive for large images or videos.”
Fast Global Image Smoothing Based on Weighted Least Squares (2014, methodology): “Thus, through this adaptive averaging process, our fast global smoother handles such noise-like structures better than the L0 smoothing, yet more efficiently.”
Let there be color! (2016, methodology): “To enable long-range propagation for image recolorization or tonal editing, various affinity-based methods have also been proposed, such as global optimization with all-pair constraints [An and Pellacini 2008; Xu et al. 2009], Radial Basis Function interpolation [Li et al.”
Semantic photo manipulation with a generative image prior (2019, background): “Example applications include color adjustment [An and Pellacini 2008; Levin et al. 2004; Reinhard et al. 2001; Xue et al. 2012], tone mapping [Durand and Dorsey 2002], image warping [Avidan and Shamir 2007], image blending [Pérez et al. 2003; Tao et al. 2010], or image reshuffling [Barnes et al.…”
Error-Tolerant Scribbles Based Interactive Image Segmentation (2014, methodology): “The scribble-based approaches have been widely used in image editing [1, 6] and image segmentation [2].”
Light Field Image Processing: An Overview (2017, background): “In the context of coarse edit propagation, Jarabo et al. [211] extended AppProp [212] to the 4 domain of light ﬁelds.”
Editing Conditional Radiance Fields (2021, background): “There has been much work on interactive tools for selecting and cloning regions [2, 35, 56, 60] and editing single still images [3, 6, 7, 34].”