Research Areas
  Computer Vision >> Fast & Robust Global Registration for Image Mosaic

CIS Distortion Analysis and its Application to Digital Image Stabilization

Ēš Image Mosaic


Ēš Sequential Global Mosaic
 In this work, to prevent large error accumulations in multiple image registration, we propose a new fast global mosaic method using sequential block matching in regularly spaced grid features. We use a specific graph structure called the sequential graph defined in each grid and propose a new shortest path search algorithm to find exact matchings in a sequential manner. Using a sequential structure, the search range to find matchings can be reduced, and the new sequential shortest path search algorithm reduces the number of block matchings.

Overall flow


Results




Ēš Practical Background
 In this work, we present a new background estimation algorithm which is able to effectively represent both background and foreground. The problem is formulated with a labeling problem over a patch-based Markov Random Field (MRF) and solved with a graph-cuts algorithm. Our method is applied to the problem of mosaic blending considering moving objects and exposure variations of rotating and zooming camera.

Basic Structure


Results




Ēš HDR Global Mosaic
 In this work, we present a global approach for constructing high dynamic range mosaics from multiple images with large exposure differences. To minimize registration errors caused by intensity mismatches in the image intensity space with low dynamic range, we propose the use of a scene radiance space with high dynamic range. By relating image intensities to scene radiances with a convenient distortion model, we robustly estimate registration parameters for the high dynamic range global mosaic, simultaneously estimating scene radiances and distortion parameters in a single framework using a computationally optimized Levenberg-Marquardt approach. Also, a simple detail-preserving contrast reduction method is proposed.

Overall flow


Image Formation Model


Results