Object-aware Gradient-Domain Image Compositing
نویسندگان
چکیده
We describe an approach to suppress bleeding artifacts without altering the boundary location in gradient-domain compositing, a technique to create seamless composites. While gradient-domain compositing has become a standard tool for many complex image editing tasks such as seamless cloning, panorama stitching or scene completion, its quality suffers from mismatches in the composited image regions. We propose an approach that is robust to nonoptimal region selection by the user without altering his selection which may be neither intended nor possible for certain compositing tasks. In addition, we present an easy-to-use extension to composite interleaving objects. The usability of our approach is demonstrated by several image compositing tasks and comparisons to current state-of-the-art algorithms in gradient-domain image compositing are presented. This is the author version of the paper. The definitive version is available at diglib.eg.org.
منابع مشابه
Nonlinear Multiresolution Image Blending
We study contrast enhancement for multiresolution image blending. In image compositing, image stitching, and image fusion, a blending operator combines coefficients of a pixel array, an image pyramid, a wavelet decomposition, or a gradient domain representation. Linear interpolation reduces variation and thereby causes contrast loss, while coefficient selection increases variation and thereby c...
متن کاملFace Replacement with HOG and SVM
This project is an attempt to completely replace faces in video frames using Support Vector Machines (SVM) as classifiers on Histogram of Oriented Gradients (HOG). We trained an SVM classifier to first identify a face in an image using HOG descriptors and then identify fiducial markers like eyes, nose and mouth in the identified face using Matlab’s Cascade Object Descriptor. Once a convex hull ...
متن کاملSeamless Image Stitching in the Gradient Domain
Image stitching is used to combine several individual images having some overlap into a composite image. The quality of image stitching is measured by the similarity of the stitched image to each of the input images, and by the visibility of the seam between the stitched images. In order to define and get the best possible stitching, we introduce several formal cost functions for the evaluation...
متن کاملImage Alignment and Stitching
Stitching multiple images together to create beautiful highresolution panoramas is one of the most popular consumer applications of image registration and blending. In this chapter, I review the motion models (geometric transformations) that underlie panoramic image stitching, discuss direct intensity-based and feature-based registration algorithms, and present global and local alignment techni...
متن کاملPerceptually Meaningful Image Editing: Depth
We introduce the concept of perceptually meaningful image editing and present two techniques for manipulating the apparent depth of objects in an image. The user loads an image, selects an object and specifies whether the object should appear closer or further away. The system automatically determines target values for the object and/or background that achieve the desired depth change. These de...
متن کامل