Global Relabeling for Continuous Optimization in Binary Image Segmentation
نویسندگان
چکیده
Recently, continuous optimization methods have become quite popular since they can deal with a variety of non-smooth convex problems. They are inherently parallel and therefore well suited for GPU implementations. Most of the continuous optimization approaches have in common that they are very fast in the beginning, but tend to get very slow as the solution gets close to the optimum. We therefore propose to apply global relabeling steps to speed up the convergence close to the optimum. The resulting primal-dual algorithm with global relabeling is applied to graph cut problems as well as to Total Variation (TV) based image segmentation. Numerical results show that the global relabeling steps significantly speed up convergence of the segmentation algorithm.
منابع مشابه
A new continuous max-flow algorithm for multiphase image segmentation using super-level set functions
We propose a graph cut based global minimization method for image segmentation by representing the segmentation label function with a series of nested binary super-level set functions. This representation enables us to use K − 1 binary functions to partition any images into K phases. Both continuous and discretized formulations will be treated. For the discrete model, we propose a new graph cut...
متن کاملEnergy-based binary segmentation of snow microtomographic images
X-ray micro-tomography has become an essential tool to investigate the mechanical and physical properties of snow, which are tied to its microstructure. A crucial step in the 3D image processing is the binary segmentation of the grayscale X-ray attenuation coefficient image to a binary ice/pore image. In the snow community, this step is usually based on global thresholding conducted with smooth...
متن کاملناحیهبندی مرز اندوکارد بطن چپ در تصاویر تشدید مغناطیسی قلبی با شدت روشنایی غیریکنواخت
The stochastic active contour scheme (STACS) is a well-known and frequently-used approach for segmentation of the endocardium boundary in cardiac magnetic resonance (CMR) images. However, it suffers significant difficulties with image inhomogeneity due to using a region-based term based on the global Gaussian probability density functions of the innerouter regions of the active ...
متن کاملLearning to approximate global shape priors for figure-ground segmentation
We present a technique for approximate minimization of two-label energy functions with higher-order or global potentials. Our method treats the energy function as a blackbox: it does not exploit knowledge of its form nor its order, as opposed to optimization schemes specialized to a particular form. The key idea is to automatically learn a lowerorder approximation of the energy function, which ...
متن کاملImage Segmentation by Branch-and-Mincut
Efficient global optimization techniques such as graph cut exist for energies corresponding to binary image segmentation from lowlevel cues. However, introducing a high-level prior such as a shape prior or a color-distribution prior into the segmentation process typically results in an energy that is much harder to optimize. The main contribution of the paper is a new global optimization framew...
متن کامل