A constraint propagation approach to structural model based image segmentation and recognition

Multi-organ localization with cascaded global-to-local regression and shape prior

“…The sequential nature of this process tends to propagate the errors and may require a backtracking procedure. To prevent such a problem, Nempont et al (2013) proposed to solve the localization and segmentation in a global fashion following a constraint network. Although this approach has the benefit of relying on strong prior knowledge, it is computationally expensive in practice (several hours) and requires a very fine parameter tuning which makes it difficult to apply in practice until now.…”

Fuzzy sets for image processing and understanding

“…These ideas were used in particular in the segmentation and recognition methods described in [11,34,36,53,85,116,153]: a concept of the ontology is used for guiding the recognition by expressing its semantic as a fuzzy set, for instance in the image domain or in an attribute domain, which can therefore be directly linked to image information.…”

“…Local fusion is often limited because spatial information is not really taken into account, and working at intermediate or higher level (for instance combining several spatial relations to guide the understanding process) is more interesting and powerful. Examples can be found in various domains [34,53,107,116,125,137,153].…”

“…To overcome the problems raised by sequential approaches, while avoiding the need of an initial segmentation, another method, still relying on a structural model, but solving the problem in a global way, was proposed in [116]. A solution is the assignment of a region of space to each model object, that satisfies the constraints expressed in the model.…”

Fuzzy methods in medical imaging

“…Another approach consists in representing all knowledge on spatial relations between structures in a graph and expressing the joint segmentation and recognition problem as a constraint satisfaction problem [47,48]. Propagators are defined for each spatial relation, and applied sequentially in order to progressively reduce the domain of each anatomical structure.…”