Mohamed Farah

A critical analysis of lifecycles and methods for ontology construction and evaluation

Evaluation is a crucial phase in ontological lifecycle, especially for ontologies that are produced using automated or semi-automated methods. In this paper, we focus on the evaluation phase in the frame of the ontology building process. We begin by a review of different ontology construction lifecycles. Next, we try to review state-of-the-art methods for ontology construction with a special interest on evaluation for assessing the quality of ontologies. Afterwards we highlight the main limits and difficulties of these methods with respect to evaluation. We conclude with a first sketch on how evaluation should be undertaken for guiding the development of high quality ontologies that are more relevant to the requirements of a particular domain.

Graph of visual words for semantic annotation of remote sensing images

Nowadays semantic image annotation is becoming more than ever a very challenging issue since it helps improving image interpretation and retrieval. Currently, most semantic annotation methods represent images as lists of keywords or histogram of visual words, and do not consider the spatial distribution of regions, nor any prior knowledge concerning objects in a scene. This obviously leads to weak and limited representation of image content. In this paper, we propose a new method for semantic image annotation that simultaneously handles all the available information of the image (contextual, spatial, and spectral). We use a remote sensing ontology as semantic resource and develop an annotation process producing a graph representing objects of a scene as well as their spatial relations.

A Semi-automatic Mapping Selection in the Ontology Alignment Process

Ontologies are considered as one of the most powerful tools for knowledge representation and reasoning. Thus, they are considered as a fundamental support for image annotation, indexing and retrieval. In order to build a remote sensing satellite image ontology that models the geographic objects that we find in a scene, their characteristics as well as their relationships, we propose to reuse existing geographic ontologies to enrich an ontological core. Reusing high quality resources (called source ontologies) helps ensuring a good quality for the extracted knowledge, and alleviating the conceptualization stage, i.e. avoiding building a new ontology from scratch. Ontology alignment is an important phase within the enrichment process. It is a process that allows discovering mappings between core and source ontologies, where each mapping is a couple of entities brought from each ontology and linked together either by an equivalence or a subsumption relationship. Such relationships are based on various similarity measures. In this paper, we first present a brief literature review of existing theoretical frameworks for similarity measures, then we describe a new alignment approach based on a semi-automatic mapping selection process that needs little human intervention. First experiments show the benefit from using the proposed approach.

Towards a new ontology matching approach based on multi-criteria analysis methods

Actually, we still have not a well established satellite image ontology that would be very useful to assist us to study major facts affecting earth, detect and monitor natural phenomena. Nevertheless, there are several domain-specific geographic ontologies that can be used as semantic resources to build such an ontology. Thus, we can start from a core satellite image ontology such as the one of Durand and enrich it using these geographic ontologies. Ontology matching is one of the principal activities in the ontology enrichment process and highly depends on the similarity measures that are considered as well as the way they are combined together in order to decide whether two concepts coming from different ontologies are alienable. In the literature, research on similarity mainly focuses on issues related to how to compute and refine similarity measures. However, few research addresses studying their dependencies and contributions in the evaluation of the overall similarity between objects to be compared. In this paper and in order to align an initial remote sensing satellite image ontology with a set of geographic ontologies, we give insights on the main similarity models as well as their associated measures. We then propose a method in order to select a reduced set of the most important similarity measures to use for the alignment. Afterwards, we present a method that can produce a ranking model that allows sorting mappings between concepts coming from two different ontologies, in a decreasing order of a global similarity score. First experimentations show that the proposed approach is promising.

Fuzzy hypergraph of concepts for semantic annotation of remotely sensed images

Annotation of images is largely studied in the literature and used in many application fields such as in image interpretation, indexation and retrieval. Manually annotating images gives valuable information on the semantic content of images, but is no longer acceptable when dealing with real corpora of images, especially in the era of big data. Content-based approaches had known great success to deal with large datasets, using low-level features such as color, texture, and shape, which are easy to compute automatically. Nonetheless, they suffer from the well known semantic gap problem, since they produce semantically very limited representations of images. In this paper, we propose a semantic image annotation approach that simultaneously handles contextual, spatial and spectral information of the image. We consider a predefined remotely sensed ontology and develop an annotation process that produces semantically rich hypergraphs representing objects in scenes, as well as their spatial and spectral attributes. We apply our approach to build a hypergraph corresponding to the Jasper Ridge AVIRIS image, showing the promising use of such representation in remote sensing.

A similarity-based framework for the alignment of an ontology for remote sensing

Building remote sensing (RS) ontologies can undoubtedly help automatic interpretation of RS images content. Ontology alignment is proven to be an effective ontology building process that enables reusing already existing semantic resources. The quality of the ontology alignment output highly depends on the similarity measures that have been considered as well as the way they are combined together. In the literature, research on similarity measures mainly focuses on how to build new or refine already existing similarity measures leading to a wide range of measures. However, few research addresses their dependencies and combination in order to evaluate the overall similarity of the concepts to be compared. In this paper, we first show how to select a reduced set of similarity measures to be used in the alignment process. Afterwards, we present a ranking model that allows sorting mappings between concepts coming from two different ontologies in a decreasing order of global similarity score. First experimentation shows that the proposed approach is promising. HighlightsOntologies can help automatic interpretation of remote sensing images content.Ontology alignment effectively helps building RS ontologies using semantic resources.The alignment process is carried out using a reduced set of similarity measures.A ranking model allows sorting mappings in a decreasing order of global similarity.