In the QBIC (Query By Image Content) project we are studying methods to query large on-line image databases using the images' content as the basis of the queries. Examples of the content we use include color, texture, and shape of image objects and regions. Potential applications include medical ("Give me other images that contain a tumor with a texture like this one"), photo-journalism ("Give me images that have blue at the top and red at the bottom"), and many others in art, fashion, cataloging, retailing, and industry. We describe a set of novel features and similarity measures allowing query by color, texture, and shape of image object. We demonstrate the effectiveness of the QBIC system with normalized precision and recall experiments on test databases containing over 1000 images and 1000 objects populated from commercially available photo clip art images, and of images of airplane silhouettes. We also consider the efficient indexing of these features, specifically addre...

We propose a method to handle approximate searching by image content in medical image databases. Image content is represented by attributed relational graphs holding features of objects and relationships between objects. The method relies on the assumption that a fixed number of "labeled" or "expected" objects (e.g., "heart", "lungs" etc.) are common in all images of a given application domain in addition to a variable number of "unexpected" or "unlabeled" objects (e.g., "tumor", "hematoma" etc.). The method can answer queries by example such as "find all X-rays that are similar to Smith's X-ray". The stored images are mapped to points in a multidimensional space and are indexed using state-of-the-art database methods (R-trees). The proposed method has several desirable properties: (a) Database search is approximate so that all images up to a prespecified degree of similarity (tolerance) are retrieved, (b) it has no "false dismissals" (i.e., all images qualifying query selection criteria are retrieved) and (c) it is much faster than sequential scanning for searching in the main memory and on the disk (i.e., by up to an order of magnitude) thus scaling-up well for large databases.

Many aspects of time-based media -- complex data encoding, compression, "quality factors," timing -- appear problematic from a data modeling standpoint. This paper proposes timed streams as the basic abstraction for modeling time-based media. Several media-independent structuring mechanisms are introduced and a data model is presented which, rather than leaving the interpretation of multimedia data to applications, addresses the complex organization and relationships present in multimedia.

We present a formal definition of the Trellis model of hypertext and describe an authoring and browsing prototype called ffTrellis that is based on the model. The Trellis model not only represents the relationships that tie individual pieces of information together into a document (i.e., the adjacencies) but specifies the browsing semantics to be associated with the hypertext as well (i.e., the manner in which the information is to be visited and presented). The model is based on Petri nets, and is a generalization of existing directed graph based forms of hypertext. The Petri net basis permits more powerful specification of what is to be displayed when a hypertext is browsed and permits application of previously-developed Petri net analysis techniques to verify properties of the hypertext. A number of useful hypertext constructs, easily described in the Trellis model, are presented. These include the synchronization of simultaneous traversals of separate paths through a hypertext, the incorporation of access controls intoahypertext (i.e., specifying nodes that can be proven to be accessible only to certain classes of browsers), and construction of multiple specialized (tailored) versions from a single hypertext.

We propose a method to handle approximate searching by image content in large image databases. Image content is represented by attributed relational graphs holding features of objects and relationships between objects. The method relies on the assumption that a fixed number of "labeled" or "expected" objects (e.g., "heart", "lungs" etc.) are common in all images of a given application domain in addition to a variable number of "unexpected" or "unlabeled" objects (e.g., "tumor", "hematoma" etc.). The method can answer queries by example such as "find all X-rays that are similar to Smith's X-ray". The stored images are mapped to points in a multidimensional space and are indexed using state-of-the-art database methods (R-trees). The proposed method has several desirable properties: (a) Database search is approximate so that all images up to a pre-specified degree of similarity (tolerance) are retrieved, (b) it has no "false dismissals" (i.e., all images qualifying query selection criteri...

As hypertext systems proliferate and as networks become more ubiquitous, an important problem is to determine how to provide hyperdocument interoperability. Instead of taking the approach of standardizing, and implementing, a large and complex set of browsing features gleaned from some "union" of existing systems, we present an approach that allows each document to encode its own browsing semantics in the links of the hyperdocument. Any implementation offering the "follow link" operation can then simulate the intent of the author. The method requires a mental shift in how a hyperdocument is viewed abstractly. Instead of treating the links of a document as defining a static directed graph, they are thought of as defining an abstract program, termed the links-automaton of the document. A branching temporal logic notation, termed HTL*, is introduced for specifying properties a document should exhibit during browsing. An automated program verification technique called model checking is ...

IN THIS thesis, a new methodology is presented which supports the efficient representation, indexing and retrieval of images by content. Images may be indexed and accessed based on spatial relationships between objects, properties of individual objects, and properties of object classes. In particular, images are first decomposed into groups of objects, called "image subsets", and are indexed by computing addresses to all such groups. All groups up to a predefined maximum size are considered. This methodology supports the efficient processing of queries by image example and avoids exhaustive searching through the entire image database.

Hypermedia systems manage interconnected information residing within a potentially wide range of data types, including text, graphics, animations, and digitized sound and images. Effective database support for hypermedia-based computing environments is essential. In order to be effective, this support must provide a variety of capabilities that are not offered by the current generation of database management systems. We report on a prototypic system called HB1 that has been designed to meet the storage needs of advanced hypermedia system architectures. HB1 is referred to as a hyperbase management system (HBMS) because it stores and manipulates information and the connectivity data that link information together to form hypermedia. HB1 is composed of three subsystems: the Object Manager (OM), Association Set Manager (ASM), and Storage Manager (SM). OM and ASM are both server processes accessible to distributed client processes via IPC interfaces. OM is an object server. ASM manages structural data applicable to the objects within OM’s repository that are involved in hypermedia connections. Physical storage is managed by SM which, in this implementation, is a semantic network database management system. HB1 instantiates a conceptual model of hypermedia that is distinctly computational, has a strong

Database indexing is a well studied problem. However, the advent of Hypertext databases opens new questions in indexing. Searches are often demarcated by pointers between text items. Thus the scope of the search may change dynamically, whereas traditional indexes cover a statically defined region such as a relation. We present techniques for indexing in hypertext databases and com-pare their performance. 1.

Abstract — In this paper, an image retrieval methodology suited for search in large collections of heterogeneous images is presented. The proposed approach employs a fully unsupervised segmentation algorithm to divide images into regions and endow the indexing and retrieval system with content-based functionalities. Low-level descriptors for the color, position, size and shape of each region are subsequently extracted. These arithmetic descriptors are automatically associated with appropriate qualitative intermediate-level descriptors, which form a simple vocabulary termed object ontology. The object ontology is used to allow the qualitative definition of the high-level concepts the user queries for (semantic objects, each represented by a keyword) and their relations in a human-centered fashion. When querying for a specific semantic object (or objects), the intermediate-level descriptor values associated with both the semantic object and all image regions in the collection are initially compared, resulting in the rejection of most image regions as irrelevant. Following that, a relevance feedback mechanism, based on support vector machines and using the low-level descriptors, is invoked to rank the remaining, potentially relevant image regions and produce the final query results. Experimental results and comparisons demonstrate in practice the effectiveness of our approach.