Image Stitching Using Structure Deformation

Abstract

Abstract-The aim of this paper is to achieve seamless image stitching without producing visual artifact caused by severe intensity discrepancy and structure misalignment, given that the input images are roughly aligned or globally registered. Our new approach is based on structure deformation and propagation for achieving the overall consistency in image structure and intensity. The new stitching algorithm, which has found applications in image compositing, image blending, and intensity correction, consists of the following main processes. Depending on the compatibility and distinctiveness of the 2D features detected in the image plane, single or double optimal partitions are computed subject to the constraints of intensity coherence and structure continuity. Afterwards, specific 1D features are detected along the computed optimal partitions from which a set of sparse deformation vectors is derived to encode 1D feature matching between the partitions. These sparse deformation cues are robustly propagated into the input images by solving the associated minimization problem in gradient domain, thus providing a uniform framework for the simultaneous alignment of image structure and intensity. We present results in general image compositing and blending in order to show the effectiveness of our method in producing seamless stitching results from complex input images.

A Multi-Resolution Blending Considering Changed Regions for Orthoimage Mosaicking

“…The method introduced several formal cost functions for the evaluation of the stitching quality in the gradient domain and defined the mosaic image as their optimum in the overlapping areas [12,13]. Jia and Tang (2008) proposed an image stitching approach by image deformation. The approach propagated the deformation into the target image smoothly and both structure deformation and colour correction were simultaneously achieved within the same framework operating in the image gradient domain [14].…”

A Real-time Stitching Algorithm for UAV Aerial Images

“…2D method is widely used in normal photo stitching because of its high speed [2]. 3D method is used in SfM(Structure from Motion) because it can handle a large number of images and get the highest accuracy [3]. However, these traditional methods cannot meet the requirements of aerial images stitching.…”

“…However, it's hard to put ground control points in the battlefield. McCartney uses a new strategy: every image set is stitched separately, then stitch all the small result images to a large one [3]. However, it will be impossible to stitch two result images if every set has a large number of images.…”

Revising Algorithm for Pathology Image Stitching

“…The subsequent variation is difficult to correct. Some scholars have pointed to the solution of such problems, for example, Jia as in [3] adopted the structured deformation approach to deal with structural dislocation, Levin as in [4] used GIST method, which is to optimize the gradient strength in the overlapping region to achieve the image mosaic. However, these methods are of limited use in solving structure dislocation issues.…”