Abstract: The present invention provides a method and system for dynamically updating descriptions of AV content, such as MPEG-7 content descriptions. One embodiment provides for a method in which first a server issues a command indicating the type of update to make to the description on a client computer. Specifically, the update may be to add, delete, or change a node in a tree or graph structure describing the AV content. For example, the command may specify an update to an instance document or its associated schema. The server also specifies the location in the description to perform the update, which may be at an absolute or relative address. Next, the server sends the data, if any, to be added or used for the change. Embodiments also provide for steps to check that the update is authorized.

Abstract: The invention relates to encoding and decoding descriptions of multimedia content in two formats using access units. The invention further relates to dynamically updating multimedia description using access units and to supporting the delivery of distributed multimedia descriptions.

Abstract: A method and apparatus for processing descriptions of audiovisual content are described. According to one embodiment, a description of audiovisual content is created, and information pertaining to reuse of the description of audiovisual content is defined. Further, the description of audiovisual content and the reuse information are stored in a repository of descriptive data to enable subsequent reuse of this description.

Abstract: The encoding and distribution of schema for multimedia content descriptions are described. Before the schema is sent to a system for use, it is encoded by assigning a first token for each component in the schema and a second token for each attribute of each component. In one aspect, a signifier is sent first to signal that the schema is to be sent in a particular format. In one aspect, the schema is sent as a whole entity first. In another aspect, some of the components of the schema are sent first, followed by the schema that refers to these components. In another aspect, the schema is sent first, followed by a set of component types. When the decoder receives the signifier and encoded schema, it looks up values for the first and second tokens in a look up table for all components and attributes to decode the schema.

Abstract: Fractal decompression and the Retinex algorithm are combined to produce a color constancy method. By using this approach, color constancy and image compression can be achieved simultaneously.

Abstract: The present invention provides improved techniques for spatial representation of data and browsing based on similarity. For example, improved techniques for spatial representation of image data and browsing the image data based on the similarities (or dissimilarities) of the images are provided. In one embodiment, a hierarchical MultiDimensional Scaling (MDS) database for a set of images is provided, which allows for computationally efficient querying and updating of an image database. In one embodiment, techniques for modifying an MDS database for images are provided to allow for more intuitive browsing (or searching) of the images.

Abstract: Updates to a content description represented as a tree are contained in a series of fragment update units. Each fragment update unit includes an update command and a navigation path, and may include a fragment payload depending on the update command. The update command is applied to a set of nodes in the tree that are selected using the navigation path. The fragment payload may contain a fragment, a fragment reference, or an attribute. The update command may be an add, delete or replace command. More than one fragment payload may be present in the fragment update unit, with each fragment payload applied to a different one of the nodes from the set selected by the navigation path.

Abstract: The content description of multimedia content contains a plurality of descriptor values that pertain to specific events within the media. A Transport Hint Table (THT) provides a set of synchronization events that indicate the relation between timed events in the media and their representation in the content's description data. The THT allows for synchronized delivery of presentation of content description data with the associated events in the media. The THT provides a set of sync events based on the existence of “time” related events obtained from the content description data, which are relevant to particular events in the media. The THT allows the transport layer encoder to use the time-related events to synchronize the scheduled delivery times of media content segments and their associated descriptions. The THT also provides information for fragmenting a description for delivery and for managing resources at the decoder.

Abstract: Updates to a content description represented as a tree are contained in a series of fragment update units. Each fragment update unit includes an update command and a navigation path, and may include a fragment payload depending on the update command. The update command is applied to a set of nodes in the tree that are selected using the navigation path. The fragment payload may contain a fragment, a fragment reference, or an attribute. The update command may be an add, delete or replace command. More than one fragment payload may be present in the fragment update unit, with each fragment payload applied to a different one of the nodes from the set selected by the navigation path.

Abstract: An apparatus for selecting digital material objects containing MPEG-7 content descriptions in response to MPEG-7 content description queries. The apparatus comprises an MPEG-7 content description query generation tool coupled with a search engine configured for searching and comparing embedded MPEG-7 META tag information within file headers, or database information thereof, to the MPEG-7 content description query. The query tool and search engine may be implemented separately or in combination. By way of example, the MPEG-7 query generation tool is implemented within an application program such as a browser, or web site, to allow the user to readily and accurately define the content being sought. The content query being sent to a search engine configured for performing MPEG-7 content searches that return hits based on content related matches.

Abstract: A method and apparatus for comparing data is described. In one embodiment, an exemplary method includes receiving a first set of data pertaining a first object and a second set of data pertaining to a second object, and comparing the first object with the second object using an earth mover's distance method that is based on computation of a series of Hausdorff distances.

Abstract: The present invention provides improved techniques for spatial representation of data and browsing based on similarity. For example, improved techniques for spatial representation of image data and browsing the image data based on the similarities (or dissimilarities) of the images are provided. In one embodiment, a hierarchical MultiDimensional Scaling (MDS) database for a set of images is provided, which allows for computationally efficient querying and updating of an image database. In one embodiment, techniques for modifying an MDS database for images are provided to allow for more intuitive browsing (or searching) of the images.

Abstract: A system and method for utilizing a graphic equalizer in performing image search procedures includes a histogram calculator for creating sets of histograms that represent corresponding images. A graphic equalizer is configured to select weighting parameters for performing the image search procedure. A feature detector calculates distance measures between histogram comparison pairs of a query image and subject images. The feature detector applies the weighting parameters to the distance measures to produce weighted distance measures. The feature detector next combines the weighted distance measures from the histogram comparison pairs to produce a composite distance measure for each of the subject images. The feature detector may then identify one or more of the subject images that are similar to the query image by analyzing the composite distance measures.

Abstract: An encoding methodology encodes an instance document describing multimedia content by determining a context node in the content description and using a schema associated with the context node to determine the maximum number of children attributes and elements of the context node. Values for required attributes and elements are encoded into required attributes and required elements sections, respectively. Values for each optional attribute and optional element present in the content description are encoded into corresponding optional attributes and optional elements sections. In one aspect, a mask is encoded that indicates which of the optional attributes or optional elements are present in the content description. In another aspect, a count of the optional attributes or optional elements is used in place of the mask. A corresponding decode methodology extracts the values of the attributes and elements from the encoded instance document to re-create the content description.

Abstract: A system and method for augmenting digital material with MPEG-7 content descriptions to enhance searching and selection of the digital material. Pursuant to the collection of content information about the digital material, either through receipt or content characterization, a set of MPEG-7 descriptor schemes and descriptors are created for constituent elements of the content. The MPEG-7 descriptions are structured as META tags, including predetermined opening and closing delimiters, which are inserted within the header field of the digital material. The MPEG-7 content description data may contain multiple content levels describing levels of associated embedded content.

Abstract: A method and apparatus for processing descriptions of audiovisual content are described. According to one embodiment, a description of audiovisual content is evaluated and a determination is made as to whether the description is an abstraction. If it is determined that the description is an abstraction, then a level of abstraction is calculated and an indicator of the level of abstraction is stored with the description of audiovisual content.

Abstract: The encoding and distribution of schema for multimedia content descriptions are described. Before the schema is sent to a system for use, it is encoded by assigning a first token for each component in the schema and a second token for each attribute of each component. In one aspect, a signifier is sent first to signal that the schema is to be sent in a particular format. In one aspect, the schema is sent as a whole entity first. In another aspect, some of the components of the schema are sent first, followed by the schema that refers to these components. In another aspect, the schema is sent first, followed by a set of component types. When the decoder receives the signifier and encoded schema, it looks up values for the first and second tokens in a look up table for all components and attributes to decode the schema.

Abstract: A system and method are disclosed for effectively utilizing universal feature detectors to generate closely-correlated universal feature information for use in various content-retrieval procedures. In certain embodiments, the universal feature detectors may generate a multi-scale gaussian pyramid based upon selected original image data by iteratively performing a series of summing transforms. The universal feature detectors may also calculate multi-scale histograms corresponding to the multi-scale gaussian pyramid, and may then perform Radon transforms upon the multi-scale histograms to produce and store corresponding Radon transform lines. The universal feature detectors may also generate a multi-scale laplacian pyramid from the multi-scale gaussian pyramid by iteratively performing a series of difference transforms. The resultant universal feature information may then advantageously provide various types of required feature information with optimal binding characteristics.

Abstract: A data document is constructed by attaching a set of child nodes, selected from nodes in an existing data document or a document fragment, at an update location, which is a node in the existing data document or the document fragment that is identified by a navigation path. When the update command is a merge, one or mode nodes (“merge nodes”) in the document fragment are identified as update locations, and the nodes in the existing data document are merged as child nodes of the merge nodes in the document fragment. When the update command is a join, a node (“join node”) in the existing data document is identified as the update location, and an ordered list specifies which nodes in the existing data document and which nodes in the document fragment are to be joined as the set of child nodes for the join node. The join and merge update commands, navigation paths, and document fragments may transmitted as a series of fragment update units to construct the data document.

Abstract: A method and an apparatus of designing a set of wavelet basis trained to fit a particular problem. The method and apparatus include constructing a neural network of arbitrary complexity using a discrete and finite Radon transform, feeding an input wavelet prototype through the neural network and its backpropagation to produce an output, and modifying the input wavelet prototype using the output.