Towards a Hybrid System Using an Ontology Enriched by Rules for the Semantic Annotation of Brain MRI Images

Abstract

To cite this version:Ammar Mechouche, Christine Golbreich, Bernard Gibaud. Towards an hybrid system using an ontology enriched by rules for the semantic annotation of brain MRI images. Abstract. This paper describes an hybrid method combining symbolic and numerical techniques for annotating brain Magnetic Resonance images. Existing automatic labelling methods are mostly statistical in nature and do not work very well in certain situations such as the presence of lesions. The goal is to assist them by a knowledge-based method. The system uses statistical method for generating a sufficient set of initial facts for fruitful reasoning. Then, the reasoning is supported by an OWL DL ontology enriched by SWRL rules. The experiments described were achieved using the KAON2 reasoner for inferring the annotations.

A Hybrid System Using Symbolic and Numeric Knowledge for the Semantic Annotation of Sulco-Gyral Anatomy in Brain MRI Images

“…Hence, it may happen in some cases that a rule expressing the propagation of a property from parts to whole cannot be fired, because an instance of Patch is defined without being connected to a known instance of gyrus by the relation partOf. KAON2 does not make all the consequences according to the first-order semantics of SWRL, but only consequences under "the DL-safe semantics" [20]. In the context of our application, using rules with ontologies enables us to meet a significant subset of our requirements.…”

“…In the context of our application, using rules with ontologies enables us to meet a significant subset of our requirements. In addition to the rules presented in the paper, standard rules may be needed for chaining ontology properties, such as the transfer of properties from parts to wholes (which are now representable in OWL 2 20 ), or dependencies in the brain cortex, for example the rule: separatesMAE(y1, x1, x2) ∧ hasSegment(y2, y1) ∧ Sulcus(y2) ∧ MAE(x1) ∧ MAE(x2) ∧ SF(y1) → separatesMAE(y2, x1, x2) allows one to propagate the separation relationship between anatomical entities from part to whole.…”

A Hybrid System for the Semantic Annotation of Sulco-Gyral Anatomy in MRI Images

“…Our brain ontology contains for each hemisphere logical definitions of 49 gyri, 5 lobes, 3 operculum, 17 gyri parts, 44 sulci, 44 sulci parts (segments), and 31 relations. The ontology is enriched by some rules increasing its expressivity (see [14] for more details).…”

“…The very large number of possible combinations led us to adopt a hybrid approach, consisting in selecting a reasonable number of hypotheses for the labelling of patches, based on an atlas, and to select the valid combinations of such hypotheses, based on existing prior knowledge about the spatial arrangement of the gyri. This part constitutes a major extension of the method presented in [14]. Section 2 provides an overview of the method and further details of the system components; section 3 describes our first experiments with both normal and pathological data; section 4 discusses these preliminary results and highlights the capabilities and current limitations of the system.…”

A Scientific Workflow Platform for Generic and Scalable Object Recognition on Medical Images

“…These approaches work well for particular tasks (e.g. [10]) but make maintenance and reusability difficult and do not scale for other object recognition tasks. Our proposed workflows have a greater maintainability since they are no longer hard-coded into the application code but are modifiable remotely, via a graphical user interface.…”