Single-Perspective Warps in Natural Image Stitching

About

Results of image stitching can be perceptually divided into single-perspective and multiple-perspective. Compared to the multiple-perspective result, the single-perspective result excels in perspective consistency but suffers from projective distortion. In this paper, we propose two single-perspective warps for natural image stitching. The first one is a parametric warp, which is a combination of the as-projective-as-possible warp and the quasi-homography warp via dual-feature. The second one is a mesh-based warp, which is determined by optimizing a total energy function that simultaneously emphasizes different characteristics of the single-perspective warp, including alignment, naturalness, distortion and saliency. A comprehensive evaluation demonstrates that the proposed warp outperforms some state-of-the-art warps, including homography, APAP, AutoStitch, SPHP and GSP.

Tianli Liao, Nan Li · 2018

Related benchmarks

Task	Dataset	Result
Image Stitching	UDIS-D (test)	mPSNR (Easy)25.82	17
Image Stitching	classical datasets	mPSNR (Easy)23.83	11
Image Stitching	Classical Datasets Easy	mPSNR23.83	9
Image Stitching	Classical Datasets Moderate	mPSNR19.16	9
Image Stitching	Classical Datasets Hard	mPSNR14.66	9
Image Stitching	Classical Datasets Average	mPSNR18.75	9
Image Stitching	UDIS-D Easy	PSNR26.98	9
Image Stitching	UDIS-D Moderate	PSNR22.67	9
Image Stitching	UDIS-D Average	PSNR21.6	9
Image Stitching	UDIS-D Hard	PSNR16.77	9