Improved Methods for Interactive Graph-Based Segmentation
Filip Malmberg, Robin Strand, Ingela Nyström, Ewert Bengtsson
Funding: TN-faculty, UU
Period: 0901-
Abstract: Image segmentation, the process of identifying and separating relevant objects and structures in an image, is a fundamental problem in image analysis. Accurate segmentation of objects of interest is often required before further processing and analysis can be performed. Despite years of active research, fully automatic segmentation of arbitrary images remains an unsolved problem.
Interactive segmentation methods use human expert knowledge as additional input, thereby making the segmentation problem more tractable. A successful semi-automatic method minimizes the required user interaction time, while maintaining tight user control to guarantee the correctness of the result. The input from the user is typically given in one of two forms:
Interactive segmentation is often phrased as an optimization problem, i.e., a solution is sought that optimizes some criterion on segmentation ``goodness'' while satisfyng the constrains provided by the user. In this project, we develop new methods for interactive segmentation, using a combinatorial approach. In 2013, results from this project were presented at the International Symposium on Mathematical Morphology (ISMM) is Uppsala.
The Stochastic Watershed
Bettina Selig, Cris Luengo, Ida-Maria Sintorn, Filip Malmberg
Funding: S-faculty, SLU
Period: 1102-
Abstract: The stochastic watershed is a method recently presented that builds on the classical seeded watershed algorithm. It creates a probability density function for edges in the image by repeated applications of the seeded watershed with random seeds. We have found that adding noise to the input image before every application of the seeded watershed greatly improves the properties of the output. These results were published this year in Pattern Recognition Letters. This year we have developed an efficient algorithm that computes the result one
would obtain after an infinite number of repetitions of the seeded watershed, and have been working towards a method to combine this algorithm with the improvements presented in our previous paper.
Adaptive Mathematical Morphology
Vladimir Curic, Cris Luengo, Gunilla Borgefors
Partner: Jesús Angulo, Centre for Mathematical Morphology, Ecole des Mines de Paris - MINES ParisTech, Fontainebleau, France;
Anders Landström, Matthew Thurley, Luleå University of Technology, Luleå;
Sébastien Lefèvre, University of South Brittany, Vannes, France;
Santiago Velasco-Forero, National University of Singapore, Republic of Singapore.
Funding: Graduate School in Mathematics and Computing (FMB)
Period: 1101-
Abstract: The construction of adaptive structuring elements that adjust their shape and size to the local structures in the image has recently been a popular topic in mathematical morphology. Despite that several methods for the construction of spatially adaptive structuring elements have been proposed, it is still an open problem, both from a theoretical and implementation point of view.
We have proposed salience adaptive structuring elements that modify their shape and size according to the saliency of the edges in the image. We have examined topological properties of salience adaptive structuring elements and investigated their applicability to image filtering. This work has been published in IEEE Journal of Selected Topics in Signal Processing. We have also proposed structuring elements with predefined shape and adaptive size based on similar type of the salience map as it was used for the construction of the salience adaptive structuring elements. Furthermore, we extended this work to salience-based parabolic structuring functions, which was presented at the International Symposium on Mathematical Morphology (ISMM'2013). More recently, we perform a comparative study of a few most important methods for constructing adaptive structuring elements as well theoretical advances how to properly define respective morphological operators. This work is currently under review.
We intend to further investigate theoretical properties of adaptive morphological operators as well as apply such operators to the task of image regularization. An extension of adaptive morphological operators towards multi-valued images and their definitions for sparse image representations are of interest in future studies.
Digital Distance Functions and Distance Transforms
Robin Strand, Gunilla Borgefors
Partner: Benedek Nagy, Dept. of Computer Science, Faculty of Informatics, University of Debrecen, Hungary; Nicols Normand, IRCCyN, University of Nantes, France
Funding: TN-faculty, UU; S-faculty, SLU
Period: 9309-
Abstract: The distance between any two grid points in a grid is defined by a distance function. In this project, weighted distances have been considered for many years. A generalization of the weighted distances is obtained by using both weights and a neighborhood sequence to define the distance function. The neighborhood sequence allows the size of the neighborhood to vary along the paths.
In 2013, papers on
The Minimum Barrier Distance
Robin Strand, Filip Malmberg
Partner: Punam K. Saha, Dept. of Electrical and Computer Engineering and the Dept. of Radiology, University of Iowa, IA, USA; Krzysztof C. Ciesielski, Dept. of Mathematics, West Virginia University, Morgantown, WV, USA; Dept. of Radiology, MIPG, University of Pennsylvania, PA, USA
Funding: TN-faculty, UU
Period: 1103-
Abstract: In this project, we introduce a distance function on a fuzzy subset that gives the minimum barrier that has to be passed to go from one point to another. Theoretical properties as well as efficient computational solutions for minimum barrier distance have been developed. An initial application of minimum barrier distance in image segmentation is presented. The experiments show that the minimum barrier distance is robust to noise and blur, and also seed point position, since it captures the total change in membership values across an interface instead of gradient as a measure of slope that is sensitive to noise and blur.
A paper on the theoretical foundation of the minimum barrier distance was published in Computer Vision and Image Understanding. Our work in this project during 2013 has been focused on finding efficient, and exact, algorithms for computing the minimum barrier distance.
Set Distances and their Application in Image Analysis
Vladimir Curic, Gunilla Borgefors
Partner: Joakim Lindblad, Nataša Sladoje, Faculty of Technical Sciences, University of Novi Sad, Serbia
Funding: Graduate School in Mathematics and Computing (FMB)
Period: 0908-
Abstract: Methods for measuring distances between sets, which is a measure of how similar the sets are, can be useful for solving various image analysis related problems, such as registration, image retrieval and segmentation evaluation. Depending on how the distance measure is defined, it exhibits different properties, such as metricity, monotonicity, continuity, sensitivity to noise, complexity and speed of computation. It is therefore of interest to study and further develop different set distance measures, to be able to select appropriate distances for the different applications. In this project, we evaluate existing and develop new set distances which are useful in image registration related problems.
We have proposed a new set distance between crisp sets of points and evaluated its usefulness for rigid body registration of binary images as well as its applicability for the real task of multi-modal 2D-3D registration of 2D histological sections of bone implant with corresponding 3D synchrotron radiation micro computed tomography (SRµCT) bone implant volumes. In addition, it has been shown that this set distance has good performances when applicable to the task of recognition of handwritten characters. This work has been accepted for publication to Pattern Analysis and Applications.
We extended our study to fuzzy objects and proposed four novel point-to-set distances defined for fuzzy or gray-level image data, two based on integration of alpha cuts and two based on the fuzzy distance transform. We further used these point-to-set distances to define distances between fuzzy sets. Theoretical study and performance evaluation of the proposed distances confirm their excellent behaviour in template matching and object classification. New distance measures enable to include and consider both spatial and intensity information, which makes them applicable to texture matching problems as well. The results of this study have been published in IEEE Transactions on Image Processing.
Direct Curvature Calculation of Surfaces in 3D Volumes
Erik Wernersson, Cris Luengo, Anders Brun, Gunilla Borgefors
Funding: S-faculty, SLU
Period: 1009 -
Abstract: Curvature is known to be a useful local descriptor of 2D surfaces, embedded in 3D space. Not only for parametric surfaces but
also estimated from objects in digital images with applications ranging from visualisation to segmentation.
Within this project, we have studied curvature calculated from the structure tensor, in contrast to the most common methods which derive curvature directly from image differentials. This opens up for new kind of processing, and especially averaging, which we hope will be of interest for the analysis of wood fibres in µCT images of paper and composite materials.