A method of embedding a high-resolution image into a large field-of-view image

Yanmei Dong*, Mingtao Pei, Yunde Jia

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

Abstract

Images captured by wide-angle cameras or fisheye cameras are with large Field-of-View (FOV) but low resolution, while images captured by conventional cameras are with high-resolution but limited FOV. To handle this contradiction, in this paper, we propose to improve the resolution of a large FOV image by embedding a high-resolution image into it. We present an image embedding method based on non-rigid transformations and mesh optimization. Without any radial distortion correction on large FOV images, we utilize non-rigid transformations to model the mathematical transformation relationship between the high-resolution image and the large FOV image. Meanwhile, we exploit mesh optimization to generate evenly distributed control points for maintaining the shape and content of the large FOV image. To verify the effectiveness of the proposed method, we capture some image pairs from a conventional camera and a large FOV camera, and exploit several classic image alignment methods to implement four baseline methods. The experimental results show that our method preserves the shape and content of the original FOV image from 58% to 82%, which significantly outperforms all baseline methods. The results indicate that the proposed method can effectively embed a high-resolution image into a large FOV image as well as preserve the shape and content of the large FOV image. In other words, the resolution of a large FOV image can be improved via high-resolution image embedding.

Original languageEnglish
Article number103912
Pages (from-to)13037-13056
Number of pages20
JournalMultimedia Tools and Applications
Volume84
Issue number14
DOIs
Publication statusPublished - Apr 2025
Externally publishedYes

Keywords

  • High-resolution image
  • Image alignment
  • Image embedding
  • Image resolution
  • Large field-of-view image

Cite this