Thinking inside the Convolution for Image Inpainting: Reconstructing Texture via Structure under Global and Local Side

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Liu, Haipeng, Wang, Yang, Qian, Biao, Rui, Yong, Wang, Meng
Format: Preprint
Veröffentlicht: 2026
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866910009922682880
author Liu, Haipeng
Wang, Yang
Qian, Biao
Rui, Yong
Wang, Meng
author_facet Liu, Haipeng
Wang, Yang
Qian, Biao
Rui, Yong
Wang, Meng
contents Image inpainting has earned substantial progress, owing to the encoder-and-decoder pipeline, which is benefited from the Convolutional Neural Networks (CNNs) with convolutional downsampling to inpaint the masked regions semantically from the known regions within the encoder, coupled with an upsampling process from the decoder for final inpainting output. Recent studies intuitively identify the high-frequency structure and low-frequency texture to be extracted by CNNs from the encoder, and subsequently for a desirable upsampling recovery. However, the existing arts inevitably overlook the information loss for both structure and texture feature maps during the convolutional downsampling process, hence suffer from a non-ideal upsampling output. In this paper, we systematically answer whether and how the structure and texture feature map can mutually help to alleviate the information loss during the convolutional downsampling. Given the structure and texture feature maps, we adopt the statistical normalization and denormalization strategy for the reconstruction guidance during the convolutional downsampling process. The extensive experimental results validate its advantages to the state-of-the-arts over the images from low-to-high resolutions including 256*256 and 512*512, especially holds by substituting all the encoders by ours. Our code is available at https://github.com/htyjers/ConvInpaint-TSGL
format Preprint
id arxiv_https___arxiv_org_abs_2602_03013
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Thinking inside the Convolution for Image Inpainting: Reconstructing Texture via Structure under Global and Local Side
Liu, Haipeng
Wang, Yang
Qian, Biao
Rui, Yong
Wang, Meng
Computer Vision and Pattern Recognition
Image inpainting has earned substantial progress, owing to the encoder-and-decoder pipeline, which is benefited from the Convolutional Neural Networks (CNNs) with convolutional downsampling to inpaint the masked regions semantically from the known regions within the encoder, coupled with an upsampling process from the decoder for final inpainting output. Recent studies intuitively identify the high-frequency structure and low-frequency texture to be extracted by CNNs from the encoder, and subsequently for a desirable upsampling recovery. However, the existing arts inevitably overlook the information loss for both structure and texture feature maps during the convolutional downsampling process, hence suffer from a non-ideal upsampling output. In this paper, we systematically answer whether and how the structure and texture feature map can mutually help to alleviate the information loss during the convolutional downsampling. Given the structure and texture feature maps, we adopt the statistical normalization and denormalization strategy for the reconstruction guidance during the convolutional downsampling process. The extensive experimental results validate its advantages to the state-of-the-arts over the images from low-to-high resolutions including 256*256 and 512*512, especially holds by substituting all the encoders by ours. Our code is available at https://github.com/htyjers/ConvInpaint-TSGL
title Thinking inside the Convolution for Image Inpainting: Reconstructing Texture via Structure under Global and Local Side
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2602.03013