Image Alignment and Stitching: A Tutorial

Abstract

This tutorial reviews image alignment and image stitching algorithms. Image alignment algorithms can discover the correspondence relationships among images with varying degrees of overlap. They are ideally suited for applications such as video stabilization, summarization, and the creation of panoramic mosaics. Image stitching algorithms take the alignment estimates produced by such registration algorithms and blend the images in a seamless manner, taking care to deal with potential problems such as blurring or ghosting caused by parallax and scene movement as well as varying image exposures. This tutorial reviews the basic motion models underlying alignment and stitching algorithms, describes effective direct (pixel-based) and feature-based alignment algorithms, and describes blending algorithms used to produce seamless mosaics. It closes with a discussion of open research problems in the area.

Registration of infrared transmission images using squared-loss mutual information

“…We took 71 images from a semi-conductor sample, which were captured in the layer-by-layer manner and small misalignments were incurred due to mechanical movements. We compared the performance of the proposed LSMI-based method with two conventional methods based on normalized cross-correlation (NCC) [3] and mutual information (MI) [16], where kernel density estimators for p(x), p(y), and p(x, y) were used when computing the gradient of MI.…”

“…On the other hand, the image registration problem we are tackling in this paper is much more challenging than the previous works because an infrared transmission image obtained from one layer is contaminated with defocused images coming from other layers. For this reason, standard linear similarity metrics such as the sum of squared differences (SSD) and the normalized cross-correlation (NCC) [1,2,3,4,5,6] are not suitable to registration of infrared transmission images.…”

Sewing Photos: Smooth Transition between Photos

“…For this case, the usually used approach is to stitch the photos [19]. However, in our case, we just need to know the affine transformation between the two photos and generate a virtual camera path crossing through the two photos by registering them first.…”

Ruler Based Automatic C-Arm Image Stitching Without Overlapping Constraint

“…These methods can be largely classified into two categories: those that require a significant modification of the C-arm machine [2] and those that try to solve the problem by purely image registration and panorama creation techniques developed in the computer vision field [3], therefore avoiding significant modification of the hardware [4][5][6][7][8]. Examples of methods in the first categories include the method introduced by Wang et al [2], where parallax-free panoramic X-ray images were generated by enabling the C-arm to rotate around its X-ray source center, relative to the patient's table.…”