Convex Quadratic Programming for Image Segmentation
Abstract. A quadratic programming formulation for multiclass image segmentation is investigated. It is proved that, in the con- vex case, the global minima of Quadratic Markov Measure Field (QMMF) models holds the non-negativity constraint. This allows one to design e ffi cient optimization algorith...
| Autor: | |
|---|---|
| Tipo de recurso: | informe técnico |
| Estado: | Versión publicada |
| Fecha de publicación: | 2009 |
| País: | México |
| Institución: | Centro de Investigación en Matemáticas |
| Repositorio: | Repositorio Institucional CIMAT |
| Idioma: | inglés |
| OAI Identifier: | oai:cimat.repositorioinstitucional.mx:1008/608 |
| Acceso en línea: | http://cimat.repositorioinstitucional.mx/jspui/handle/1008/608 |
| Access Level: | acceso abierto |
| Palabra clave: | info:eu-repo/classification/MSC/Segmentación de Imágenes info:eu-repo/classification/cti/1 info:eu-repo/classification/cti/12 info:eu-repo/classification/cti/1203 info:eu-repo/classification/cti/120302 |
| Sumario: | Abstract. A quadratic programming formulation for multiclass image segmentation is investigated. It is proved that, in the con- vex case, the global minima of Quadratic Markov Measure Field (QMMF) models holds the non-negativity constraint. This allows one to design e ffi cient optimization algorithms. We also proposed a (free parameter) inter–pixel a ffi nity measure more related with the classes memberships than with color or gray gradient based standard methods. Moreover, it is introduced a formulation for computing the pixel likelihoods by taking into account local con- text and texture properties. We demonstrate the QMMFs capabil- ities by experiments and numerical comparisons with interactive two-class segmentation as well as in the simultaneous estimation of segmentation and (parametric and non-parametric) generative models. |
|---|