EXPLOITING HIGHER ORDER UNCERTAINTY IN IMAGE ANALYSIS

Soft computing is a group of methodologies that works synergistically to provide flexible information processing capability for handling real-life ambiguous situations. Its aim is to exploit the tolerance for imprecision, uncertainty, approximate reasoning, and partial truth in order to achieve tractability, robustness, and low-cost solutions. Soft computing methodologies (involving fuzzy sets, neural networks, genetic algorithms, and rough sets) have been successfully employed in various image processing tasks including image segmentation, enhancement and classification, both individually or in combination with other soft computing techniques. The reason of such success has its motivation in the fact that soft computing techniques provide a powerful tools to describe uncertainty, naturally embedded in images, which can be exploited in various image processing tasks. The main contribution of this thesis is to present tools for handling uncertainty by means of a rough-fuzzy framework for exploiting feature level uncertainty. The first contribution is the definition of a general framework based on the hybridization of rough and fuzzy sets, along with a new operator called RF-product, as an effective solution to some problems in image analysis. The second and third contributions are devoted to prove the effectiveness of the proposed framework, by presenting a compression method based on vector quantization and its compression capabilities and an HSV color image segmentation technique.