Desenvolvimento de um método híbrido para segmentação imagens utilizando crescimento de regiões e lógica fuzzy

Image segmentation is a process that groups pixels related to one or more local properties obtained from the values of the pixels or numerical results performed on a neighborhood. It is considered as a critical step in the system of image processing, when considering systems that uses classical logic. Errors committed during this step can influence so undesirable the results produced in further analysis. One way to reduce this problem is to introduce the fuzzy paradigm in segmentation techniques, so the appropriate representation of uncertainty allows more information to be maintained throughout the process producing a result with a degree of uncertainty. This article presents a hybrid segmentation method based on region growing technique using classical logic to determine an original region of interest and fuzzy logic to make general segmentation of all features that resemble the original region of interest according to a criterion of similarity predetermined by a membership function.Pages: 7404-741