Higher-order regularization and morphological techniques for image segmentation
Image segmentation is an important ﬁeld in computer vision and one of its most active research areas, with applications in image understanding, object detection, face recognition, video surveillance or medical image processing. Image segmentation is a challenging problem in general, but especially in the biological and medical image ﬁelds, where the imaging techniques usually produce cluttered and noisy images and near-perfect accuracy is required in many cases.
In this thesis we ﬁrst review and compare some standard techniques widely used for medical image segmentation. These techniques use pixel-wise classiﬁers and introduce weak pair-wise regularization which is insufﬁcient in many cases. We study their difﬁculties to capture higher-level structural information about the objects to segment. This deﬁciency leads to many erroneous detections, ragged boundaries, incorrect topological conﬁgurations and wrong shapes. To deal with these problems, we propose a new regularization method that learns shape and topological information from training data in a non-parametric way using higher-order potentials.
KeywordsImage Segmentation, Random Fields, Higher-order Models, Geodesic Active Contours, Mathematical Morphology
Copyright (c) 2015 Pablo Márquez Neila
This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.