Content uploaded by Mara Abel
Author content
All content in this area was uploaded by Mara Abel
Content may be subject to copyright.
A preview of the PDF is not available
Ontology-based annotation of atomic and abstract petrographic image features
Abstract and Figures
Ontology-based annotation of images is intended to allow standard and explicit meanings to be assigned to objects and features that an observer perceives there. In practice, interpretation of images in sedimentary petrology dealing with atomic items by themselves in an ontology (such as concepts, attributes and values) is insufficient to capture higher-level semantics (e.g. evidence, findings, justifications). As a response to that problem, the paper proposes two conceptual annotation levels: atomic, and abstracted. We describe here a knowledge-based model that helps users to make their image-interpreting knowledge explicit and apply in interpretation of oil-reservoir rock images. Two concepts are specific to this work: visual chunks, which represent visual patterns of domain terms, and K-graphs, which are used to state knowledge-intensive connections between items of observed evidence and interpretations for this evidence. The essential details of the framework are not at all constrained by the geological application: they can be applied without change in any other subject-areas where reasoning and annotation of images are the primary requirement.
Figures - uploaded by Mara Abel
Author content
All figure content in this area was uploaded by Mara Abel
Content may be subject to copyright.
... [16] Existing ontologies are often used in medical systems, e.g., cell image analysis ontology, [17] ontology-driven image analysis for histopathological images, [18] ontology of mammography, [19] or geoinformatics/geodesy, e.g., ontology-driven geographic information systems. [20,21] The usefulness of ontologies for classifying and indexing images is also noted in systems such as semantic image annotation, [22] PetroGrapher, [23] annotation of image segments with ontologies, [24] or satellite image search and sorting system. [25] Ontology in image analysis systems allows to: • Obtain detailed and precise classification results [26,27] • Easily extend to other components [20] • Use existing ontologies and adapt them to the requirements of new ontologies [17] • Obtain possibilities of data mining [28] • Apply fuzzy logic to uncertain data. ...
Knowledge-based systems have become ubiquitous in recent years. Knowledge-base developers need to be able to share and reuse knowledge bases that they build. Therefore, interoperability among different knowledge-representation systems is essential. The Open Knowledge-Base Connectivity protocol (OKBC) is a common query and construction interface for frame-based systems that facilitates this interoperability. Protege-2000 is an OKBC-compatible knowledge-base-editing environment developed in our laboratory. We describe Protege-2000 knowledge model that makes the import and export of knowledge bases from and to other knowledge-base servers easy. We discuss how the requirements of being a usable and configurable knowledge-acquisition tool affected our decisions in the knowledge-model design. Protege-2000 also has a flexible metaclass architecture which provides configurable templates for new classes in the knowledge base. The use of metaclasses makes Protege-2000 easily extensible and enables its use with other knowledge models. We demonstrate that we can resolve many of the differences between the knowledge models of Protege-2000 and Resource Description Framework (RDF)-a system for annotating Web pages with knowledge elements-by defining a new metaclass set. Resolving the differences between the knowledge models in declarative way enables easy adaptation of Protege-2000 as an editor for other knowledge-representation systems.
MDA-903-90-K-0112. The Institute for the Learning Sciences was established in 1989 with Case-based reasoning is the technique of solving new problems by adapting solutions that were used to solve old problems. This reliance on previous experiences (or cases) is a hallmark of case-based reasoning. Each case can contain a great deal of information including a description of the situation that was encountered, ways in which the situation di ered from similar situations, and how the system reacted to the situation. In this report we trace the development of the case-based reasoning (CBR) paradigm and discuss in more detail the advantages of CBR as a problem-solving methodology. We then describe the general CBR algorithm and some of the fundamental issues that must be dealt with in any CBR system. Next, we present a survey of CBR systems that have been built to perform various tasks along with pointers for further reading. Finally, we conclude with a short discussion of the implications of CBR as a cognitive model and some pointers on how to go about building a CBR system. Contents i Case-Based Reasoning 1 1
This paper suggests a theoretical basis for identifying and classifying the kinds of subjects a picture may have, using previously developed principles of cataloging and classification, and concepts taken from the philosophy of art, from meaning in language, and from visual perception. The purpose of developing this theoretical basis is to provide the reader with a means for evaluating, adapting, and applying presently existing indexing languages, or for devising new languages for pictorial materials; this paper does not attempt to invent or prescribe a particular indexing language.
The determination of detrital modes in graywackes and arkoses, here grouped as subquartzose sandstones, requires special attention to the identification of detrital grain types, as defined operationally, and to the recognition of detrital textures, as opposed to textural elements of diagenetic origin. Accurate detrital modes can yield specific information on provenance that can be gained in no other way. Reserving the terms graywacke and arkose for imprecise field descriptions and abandoning the proportion of matrix as a prime means of classification, subquartzose sandstones can be described adequately using six numerical grain parameters. Three primary parameters summing to 100 permit quantitative serial designation of rock types in the form QxFyLz, w ere Q is total quartzose grains, F is total feldspar grains, and L is total unstable lithic fragments in the framework. Three secondary parameters in the form of ratios permit desirable refinement within each primary parameter: C/Q where C is total polycrystalline quartzose grains, P/F where P is total plagioclase grains, and V/L where V is total volcanic lithic grains. Stable grains, whose sum is the parameter Q, are essentially pure silica in mineralogy, and include both monacrystalline quartz and polycrystalline lithic fragments. The latter are gradational to less quartzose types and difficult to distinguish from felsite grains, which display relict textures of volcanic origin and internal relief owing to their polyminerallic nature. Lithic fragments are subdivided at two levels into our main categories within which are various subcategories: (a) volcanic fragments include felsitic, microlitic, lathwork, and vitric types; (b) clastic fragments include silty-sandy and argillaceous types; (c) tectonite fragments include metasedimentary quartzose types and metavolcanic feldspathic or ferromagnesian types; and (d) microgranular fragments include hypabyssal, hornfelsic, and indurated sedimentary types. Interstitial materials include (a) exotic cements like calcite or zeolite; (b) homogeneous, monominerallic phyllosilicate cement displaying textures indicative of porefilling; (c) clayey detrital lutum called protomatrix; (d) recrystallized lutum or protomatrix called orthomatrix with relict detrital texture| (e) murky, polyminerallic, diagenetic pore-filling called epimatr x, whose growth is accompanied by alteration of framework grains; and (f) deformed and recrystallized lithic fragments called pseudomatrix. Extensive albitization and other alterations complicate the interpretation of detrital modes in many rocks, but potential errors commonly can be avoided by close attention to mineralogy and relict texture except where tectonite fabrics disrupt or transpose the original detrital framework. Most subquartzose sandstones were derived from one. or a mixture, of three salient provenance types: (a) volcanic terranes, yielding feldspatholithic rocks nearly free of quartz, (b) plutonic terranes, yielding feldspathic rocks with few lithic fragments, and (c) uplifted sedimentary and metasedimentary "tectonic" terranes, yielding "chert-grain" lithic rocks with f w quartz and feldspar grains except where recycled volcanic or plutonic detritus is abundant. Many voluminous accumulations of subquartzose sandstones near continental margins had as their provenance complex volcano-plutonic orogens representing ancient magmatic arcs analogous to modern arcs associated with trenches.
