Image Completion Approaches Using the Statistics of Similar Patches

By Kaiming He. 2014

169 citations recorded by the source corpus.

Image completion involves filling missing parts in images. In this paper we address this problem through novel statistics of similar patches. We observe that if we match similar patches in the image and obtain their offsets (relative positions), the statistics of these offsets are sparsely distributed. We further observe that a few dominant offsets provide reliable information for completing the image. Such statistics can be incorporated into both matching-based and graph-based methods for image completion. Experiments show that our method yields better results in various challenging cases, and is faster than existing state-of-the-art methods.

Subfields: Computer Vision and Pattern Recognition, Generative Models

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 Generative Models

Citations in the corpus

Image Inpainting via Generative Multi-column Convolutional Neural Networks (2018, methodology): “To preserve structure, patch priority computation speciﬁes the patch ﬁlling order [4, 7, 8, 21].”
Missing Data Reconstruction in Remote Sensing Image With a Unified Spatial–Temporal–Spectral Deep Convolutional Neural Network (2018, methodology): “based methods [17]–[19], partial differential equation (PDE)based methods [20], [21], variational methods [22], [23], and”
Flow-edge Guided Video Completion (2020, methodology): “These methods find correspondences for content transfer either via patch-based synthesis [1, 39] or by solving a labeling problem with graph cuts [12, 32].”
3D Photography Using Context-Aware Layered Depth Inpainting (2020, methodology): “Inspired by the success of texture synthesis [9, 8], examplebased methods complete the missing regions by transferring the contents from the known regions of the image, either through non-parametric patch-based synthesis [58, 1, 5, 20] or solving a Markov Random Field model using belief propagation ”
Text-Guided Neural Image Inpainting (2020, background): “Image inpainting Traditional diffusion-based or patch-based approaches [27, 31] fill missing regions by propagating neighbor information[2, 37] or copying from similar patches [8, 36] of the background based on low-level features.”
Robust Image Completion via Deep Feature Transformations (2019, unknown): “In general, existing work on image inpainting can be categorized into a number of classes, including interpolation methods [31], [32], exemplar based methods [33], partial differential equation (PDE) based methods [34], variational methods [35], and learning-based methods [25], [36].”