Progressive Image Inpainting with Full-Resolution Residual Network

By Zongyu Guo, Zhibo Chen, Tao Yu, Jiale Chen, Sen Liu. 2019

118 citations recorded by the source corpus.

Recently, learning-based algorithms for image inpainting achieve remarkable progress dealing with squared or irregular holes. However, they fail to generate plausible textures inside damaged area because there lacks surrounding information. A progressive inpainting approach would be advantageous for eliminating central blurriness, i.e., restoring well and then updating masks. In this paper, we propose full-resolution residual network (FRRN) to fill irregular holes, which is proved to be effective for progressive image inpainting. We show that well-designed residual architecture facilitates feature integration and texture prediction. Additionally, to guarantee completion quality during progressive inpainting, we adopt N Blocks, One Dilation strategy, which assigns several residual blocks for one dilation step. Correspondingly, a step loss function is applied to improve the performance of intermediate restorations. The experimental results demonstrate that the proposed FRRN framework for image inpainting is much better than previous methods both quantitatively and qualitatively.

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

Resolution-robust Large Mask Inpainting with Fourier Convolutions (2021, neutral): “Another trend is progressive approaches [62, 12, 25, 61].”
Deep High-Resolution Representation Learning for Visual Recognition (2019, background): “There are already followup works, e.g., image stylization [83], inpainting [50], image enhancement [62], image dehazing [1], temporal pose estimation [6], and drone object detection [190].”
Image Inpainting: A Review (2019, methodology): “for example, the proposed architectures in [66,68,69] use the same maks for damaging the images before training their models.”
Recurrent Feature Reasoning for Image Inpainting (2020, background): “An alternative scheme is to inpaint in a progressive fashion from the hole boundary to the center [31, 9].”
Image Inpainting with Learnable Feature Imputation (2020, methodology): “2 To prevent ourselves from cherry-picking qualitative examples, we present several images (with corresponding masks) chosen by previous state-of-the-art papers [11,41,48,52], thus copying their selection.”
MAT: Mask-Aware Transformer for Large Hole Image Inpainting (2022, background): “To allow for high-resolution image inpainting, a few attempts have been made to develop progressively generation systems, such as [18, 32, 41, 71, 72].”
High-Resolution Image Inpainting with Iterative Confidence Feedback and Guided Upsampling (2020, methodology): “Guo et al. [17] propose to gradually ﬁll a hole using consecutive residual blocks.”