English

PairingNet: A Learning-based Pair-searching and -matching Network for Image Fragments

Computer Vision and Pattern Recognition 2025-01-23 v2 Graphics

Abstract

In this paper, we propose a learning-based image fragment pair-searching and -matching approach to solve the challenging restoration problem. Existing works use rule-based methods to match similar contour shapes or textures, which are always difficult to tune hyperparameters for extensive data and computationally time-consuming. Therefore, we propose a neural network that can effectively utilize neighbor textures with contour shape information to fundamentally improve performance. First, we employ a graph-based network to extract the local contour and texture features of fragments. Then, for the pair-searching task, we adopt a linear transformer-based module to integrate these local features and use contrastive loss to encode the global features of each fragment. For the pair-matching task, we design a weighted fusion module to dynamically fuse extracted local contour and texture features, and formulate a similarity matrix for each pair of fragments to calculate the matching score and infer the adjacent segment of contours. To faithfully evaluate our proposed network, we created a new image fragment dataset through an algorithm we designed that tears complete images into irregular fragments. The experimental results show that our proposed network achieves excellent pair-searching accuracy, reduces matching errors, and significantly reduces computational time. Details, sourcecode, and data are available in our supplementary material.

Keywords

Cite

@article{arxiv.2312.08704,
  title  = {PairingNet: A Learning-based Pair-searching and -matching Network for Image Fragments},
  author = {Rixin Zhou and Ding Xia and Yi Zhang and Honglin Pang and Xi Yang and Chuntao Li},
  journal= {arXiv preprint arXiv:2312.08704},
  year   = {2025}
}

Comments

25 pages, 19 figures, 6 tables

R2 v1 2026-06-28T13:50:33.749Z