Deformable models

We have developed methodologies for both 2D and 3D segmentation and reconstruction based on deformable models. The proposed models integrate features of region based and boundary based segmentation methods in order to fit the contours of the objects and model its inner topology. Also, they implement automatic procedures, the so called topological changes, that alter the mesh structure and allow the segmentation of complex features such as pronounced curvatures or holes, as well as the detection of several objects in the scene. This is one of the main advantages of the proposed models.
The evolution of the deformable model is governed by energy functions defined in such a way that the function reaches a minimum near the features of interest. Hence, the deformation process is based on an energy minimization process. In this sense, we have tested our proposals with well-known optimisation procedures. Moreover, we have analyzed evolutionary techniques, such as genetic algorithms, in order to increase the accuracy of the results.

Main publications

  • N. Barreira, M. G. Penedo, L. D. Cohen, M. Ortega, «Topological Active Volumes: a Topology-Adaptive Deformable Model for Volume Segmentation», Pattern Recognition, 43 (1), 255-266, 2010.
Abstract
This paper proposes a generic methodology for segmentation and reconstruction of volumetric datasets based on a deformable model, the Topological Active Volumes (TAV). This model integrates features of region based and boundary based segmentation methods in order to fit the contours of the objects and model its inner topology. Also, it implements automatic procedures, the so called topological changes, that alter the mesh structure and allow the segmentation of complex features such as pronounced curvatures or holes, as well as the detection of several objects in the scene. This work presents the TAV model and the segmentation methodology and explains how the changes in the TAV structure can improve the adjustment process. Also, the accuracy of proposed methodology is proved with several synthetic and real images.
  • O. Ibáñez, N. Barreira, J. Santos, M. G. Penedo, «Genetic Approaches for Topological Active Nets Optimization», Pattern Recognition, 42, 907-917, 2009.
Abstract
This paper presents two evolutionary approaches to the energy minimization prob- lem of the Topological Active Net model. This is a deformable model used for image segmentation that fits not only the surfaces but also the inner side of the objects. It detects concave and convex surfaces, holes, or several objects in the scene by means of topological changes in its structure. The first evolutionary approach is a genetic algorithm with adapted operators whereas the second one is a hybrid combination of this algorithm with a greedy one. Both genetic approaches improve the accuracy of the segmentation results but only the hybrid one is able to perform topological changes.
  • J. Novo, M. G. Penedo, J. Santos, «Localisation of the Optic disc by means of GA-Optimised Topological Active Nets», Image and Vision Computing, 27, 1572-1584, 2009.
Abstract
In this paper we propose a new approach to the optic disc segmentation process in digital retinal images by means of Topological Active Nets (TAN). This is a deformable model used for image segmentation that integrates features of region-based and edge-based segmentation techniques, being able to fit the edges of the objects and model their inner topology. In this paper the active nets incorporate new energy terms for the optic disc segmentation and their optimisation is performed with a genetic algorithm, with adapted or new ad hoc genetic operators. There is no need of any pre-processing of the images, with a simultaneous localisation and segmentation of the optic disc. We present representative results of optic disc segmentations showing the advantages of the approach, with images focusing on the optic disc or on the macula, and with images with different levels of noise and lesion areas.
  • N. Barreira, M. G. Penedo, «Topological Active Volumes», Advances in Intelligent Vision Systems: Methods and Applications; EURASIP Journal on Applied Signal Processing, 13(1), 1939-1947, 2005.
Abstract
The Topological Active Volumes (TAV) model is a general model for 3D image segmentation. It is based on deformable models and integrates features of region-based and boundary-based segmentation techniques. Besides segmentation, it can also be used for surface reconstruction and topological analysis of the inside of detected objects. The TAV structure is flexible and allows topological changes in order to improve the adjustment to object’s local characteristics, find several objects in the scene and identify and delimit holes in detected structures. This paper describes the main features of the TAV model and shows its ability to segment volumes in an automated manner.