Topology-preserving perceptual segmentation using the Combinatorial Pyramid
Scene understanding and other high-level visual tasks usually rely on segmenting the captured images for dealing with a more efficient mid-level representation. Although this segmentation stage will consider topological constraints for the set of obtained regions (e.g., their internal connectivity),...
| Autores: | , , |
|---|---|
| Tipo de recurso: | artículo |
| Estado: | Versión publicada |
| Fecha de publicación: | 2010 |
| País: | España |
| Institución: | Universidad de Sevilla (US) |
| Repositorio: | idUS. Depósito de Investigación de la Universidad de Sevilla |
| OAI Identifier: | oai:idus.us.es:11441/26205 |
| Acceso en línea: | http://hdl.handle.net/11441/26205 |
| Access Level: | acceso abierto |
| Palabra clave: | Topology-preserving image segmentation Perceptual grouping Irregular pyramids Combinatorial pyramids |
| Sumario: | Scene understanding and other high-level visual tasks usually rely on segmenting the captured images for dealing with a more efficient mid-level representation. Although this segmentation stage will consider topological constraints for the set of obtained regions (e.g., their internal connectivity), it is typical that the importance of preserving the topological relationships among regions will be not taken into account. Contrary to other similar approaches, this paper presents a bottom-up approach for perceptual segmentation of natural images which preserves the topology of the image. The segmentation algorithm consists of two consecutive stages: firstly, the input image is partitioned into a set of blobs of uniform colour (pre-segmentation stage) and then, using a more complex distance which integrates edge and region descriptors, these blobs are hierarchically merged (perceptual grouping). Both stages are addressed using the Combinatorial Pyramid, a hierarchical structure which can correctly encode relationships among image regions at upper levels. The performance of the proposed approach has been initially evaluated with respect to groundtruth segmentation data using the Berkeley Segmentation Dataset and Benchmark. Although additional descriptors must be added to deal with small regions and textured surfaces, experimental results reveal that the proposed perceptual grouping provides satisfactory scores. |
|---|