| Summary: | Image segmentation has emerged as one of the main research interests in computer vision over the past decade, which involve many applications, such as medical imaging and biometric identifications. It can be defined as a process of partitioning part of, the whole of, or objects within an image into meaningful regions for further analysis. Image segmentation is divided into variational and non-variational models. Non-variational models are highly dependent on large amounts of data and labels that are not always available and which has led some researchers to focus more on variational models. The most efficient variational model is the active contour model (ACM). However, the main problem faced by variational ACMs is segmenting intensity inhomogeneity images. Hence, this paper presents some strategies that have been utilised in modelling variational ACMs to overcome the problem. One recent strategy is integrating fuzzy theory to help with segmentation. Thus, this paper also highlights a recent type of variational ACM, selective-based variational ACM, that incorporates fuzzy theory in its formulation and discusses the advantages and the challenges of the segmentation model. Finally, several issues and future research directions in variational fuzzy region-based ACM are discussed. © 2023 IEEE.
|
|---|