CN103699572A - Digital media content and resource integration and sharing method in cloud environment - Google Patents

Abstract

Translated fromChinese

本发明涉及一种云环境下数字媒体内容资源的集成共享方法，主要包括：本发明针对数字媒体资源分布相对独立、分散和封闭建立一种统一视图，对数字媒体数据进行集成与共享；以WordNet为本体蓝本，构造了不同领域的可扩展的数字媒体资源的本体；建立一种数字媒体语义的表达方法，支持对数字媒体云素材进行不同粒度下的标注，对数字媒体云素材进行统一管理和使用。本发明整合异地的数字媒体资源和数据库，实现对数字媒体资源有效的共享、组织、管理和处理。支持不同粒度的数字媒体素材语义标注和检索。

The invention relates to a method for integrating and sharing digital media content resources in a cloud environment, mainly comprising: the invention establishes a unified view for relatively independent, decentralized and closed distribution of digital media resources, and integrates and shares digital media data; Based on the ontology blueprint, the ontology of scalable digital media resources in different fields is constructed; an expression method of digital media semantics is established to support the labeling of digital media cloud materials at different granularities, and the unified management and integration of digital media cloud materials use. The invention integrates digital media resources and databases in different places, and realizes effective sharing, organization, management and processing of digital media resources. Supports semantic annotation and retrieval of digital media materials at different granularities.

Description

A kind of integrated shared method of digital media content resource under cloud environment

Technical field

The present invention relates to virtual reality, knowledge representation and database technical field, be specifically related to the integrated shared method of digital media content resource under a kind of cloud environment, specifically integrated shared, the material database of Facing Digital media cloud material are constructed and Digital Media semantic meaning representation method.

Background technology

Digital Media is an integrated technology interdisciplinary, it is taken as the leading factor with information science and digital technology, take mass media theory as foundation, take modern art as guidance, information dissemination technology is applied to culture, art, business, education and management field, comprise graph and image processing technology, human-computer interaction technology, virtual reality technology and Mass media technology etc., there is obviously multidisciplinary, multi-field, multidirectional technology cross characteristic.

The development of digital media service and application need the support of all kinds of digital media resources.Digital media resource be growing amount large, use extensively, complex structure, researching value are high, quantity of information rich in natural resources kenel, comprise picture resource, video material, three-dimensional model, audio resource, software resource, data set resource etc.But, on the one hand, current digital media resource quantity is magnanimity and increases, resource is relatively independent again, disperse and sealing, do not realize effectively fully and sharing, cause information redundancy, exist utilization factor low, storage resources waste, the situations such as resource duplicate construction, become the bottleneck problem of numerous digital media applications systems, make content production and the service of Digital Media can not make full use of existing affluent resources, had a strong impact on the development of China Digital Media, therefore must set up unified Digital Media materials database, effectively integrate digital media resource, to unify, effectively management.On the other hand, Digital Media material has a large amount of retrievals and application demand, and it is very outstanding that the efficiency of data acquisition and accuracy problem become.How effectively according to the characteristic of digital media data, to go these data of access, make people can from a large amount of digital media informations, find easily own interested associated clip to become a kind of urgent demand.User is natural language description that contain in resource, available what carry out that digital media resource when retrieval pay close attention to, be easy to the semantic information that user understands and represents, and tradition inquiry is to inquire about on Symbol matching level, there is no on the one hand the query intention of abundant extending user, also ignored the abundant semantic information of Digital Media material simultaneously, and semantic query can fundamentally improve information retrieval performance, the effective knowledge semantic description that this just need to provide the digital media resource of magnanimity under cloud environment, marks and expresses resource.

Comprehensive above-mentioned 2 points, the building method of the semantic material database of a set of Facing Digital media resource of needs.

Summary of the invention

According to the actual requirements and key issue, the object of the invention is to: the semantic material database building method that proposes Facing Digital media cloud material.By integratedly sharing realizing digital media data, the body of the digital media resource of structure different field and realize Digital Media cloud material is carried out to unified management and use by setting up a kind of semantic meaning representation method to Digital Media, and realize semantic accurately retrieval.

The technical solution used in the present invention is: a kind of integrated shared method of digital media content resource under cloud environment, is characterized in that following steps:

Step (1), definition accessing data base interface, utilize Heterogeneous Digital media database access and integration middleware under cloud environment, and the original figure media materials of different regions is carried out to unified management;

Ontological concept classification and the expression of step (2), proposition Digital Media material, take WordNet as body source, built a Digital Media material Ontology storehouse for different application field;

Step (3), by Ontology tree node and original figure media materials, set up corresponding relation, a kind of standardization semantic description method is provided, realize to original figure media materials the semantic tagger under different grain size.

The digital media resource of the isomery that the wide area that comprises image resource, video resource, three-dimensional model resource, audio resource, data set resource and software resource that wherein, the digital media data storehouse access and integration middleware in step (1) is integrated distributes.

Wherein, in step (1), definition accessing data base interface are to set up a kind of unified mysql database format, utilize Heterogeneous Digital media database access and integration middleware under cloud environment, at the raw data word bank in digital media data storehouse, preserve the copy in the digital media data storehouse of isomery, and utilize the interface of middleware and the digital media data storehouse of isomery to keep synchronizeing renewal, can guarantee the consistance of data layout like this, avoid too much data-interface and data-switching, to improve search efficiency.

Wherein, the level (upper/the next relation) that passes through to extract the implication relation in wordnet of the Digital Media material in step (2), form the Top-level Ontology in Ontology storehouse, and enrich Digital Media interdependent node by researching and analysing the semantic feature of mass digital media resource, set up domain body.Final ontology information is to deposit in Ontology storehouse with tree structure, has set up the Ontology tree of a complete field of digital media, and body has versatility, completeness and the extensibility of Facing Digital field of media.

Wherein, semantic tagger under different grain size in step (3) refers to label for labelling and object outline mark, utilize semantic tagger instrument, object in media materials, action and Scene Semantics are marked, thereby set up digital media resource to the mapping relations of domain body, and then effectively according to Ontology, remove management and retrieval digital media data.

Principle of the present invention is:

First, the present invention is directed to distribute relatively independent, dispersion and sealing of digital media resource and set up a kind of unified view, definition accessing data base interface, build Heterogeneous Digital media database access and integration middleware under cloud environment, set up a kind of unified mysql database format and at the raw data word bank in digital media data storehouse, preserve the copy in the digital media data storehouse of isomery, original figure media materials to different regions carries out unified management, and provides unified view and interface to digital media applications developer and final user.

Secondly, the present invention has built a Digital Media material Ontology storehouse for different application field.This semantic base be take WordNet as body source, the psychologist of WordNet Shi You Princeton university, and the meaning of pressing word of linguist and Computer Engineer's co-design forms one " network of word ".The present invention is from the abstract concept of top layer of WordNet, things (thing) starts, by extracting the level (upper/the next relation) of implication relation in WordNet, set up the Top-level Ontology in Ontology storehouse, and enrich Digital Media interdependent node and set up domain body by researching and analysing the semantic feature of mass digital media resource.Finally set up an Ontology tree, wherein each node represents the concept that a class is general.

Finally, the present invention sets up corresponding relation by Ontology tree node and original figure media materials, and a kind of standardization semantic description method is provided, and realizes the semantic tagger to original media material.By the content information of Digital Media material Ontology Language formalization representation, support varigrained semantic tagger simultaneously.Realization is to the management of body and the semantic retrieval to material.

Compare with other video image material databases of prior art, feature of the present invention is:

(1), the present invention integrates digital media resource and the database in strange land, realize to digital media resource effectively share, tissue, management and process.

(2), the present invention has extendible Domain and ontology knowledge.

(3), the present invention supports varigrained Digital Media material semantic tagger and retrieval.The material database that the present invention constructs not only can mark and retrieve complete media materials material, also the object in media materials, action and Scene Semantics are marked or retrieved, and provide unified view and interface to digital media applications developer and final user.

(4), the present invention marks service and can coordinate user's incentive system, encourage user to share Digital Media media resource and add mark to resource in storehouse.Improve the degree of share of resource, creating knowledge resource (as academic data set) etc.

Accompanying drawing explanation

Fig. 1 is general structure schematic diagram of the present invention (knowledge representation model);

Fig. 2 is Ontology view of the present invention;

Fig. 3 is the label for labelling of digital media resource of the present invention;

Fig. 4 is the object outline mark of image resource;

Fig. 5 is database structure schematic diagram of the present invention;

Fig. 6 is the material retrieval flow figure of Digital Media of the present invention;

Fig. 7 is the material semantic retrieval result figure of Digital Media of the present invention.

Embodiment

Below in conjunction with accompanying drawing and embodiment, the present invention is elaborated.

Consult Fig. 1 general structure schematic diagram of the present invention.Digital media resource cloud is by database unified view, resource registering module, and semantic tagger module and resource retrieval module form.Digital media resource cloud database unified view is comprised of materials database, Ontology storehouse and semantic marker storehouse, for storing, manage in different local original figure media data base resource, knowledge base resource and signature library resources.The present invention has set up a kind of unified mysql database format, utilize Heterogeneous Digital media database access and integration middleware under cloud environment, by in the original figure media materials data importing materials database in the digital media data storehouse of isomery, the Ontological concept in the digital media data storehouse of isomery is incorporated in Ontology storehouse.Utilize the interface of middleware, in materials database, to the modification of media materials essential information, deletion, can synchronously be updated in the digital media data storehouse of isomery, guarantee the consistance of data.Effectively the digital media resource of isomery is carried out to unified management.Materials database is for directly storing the essential information of various digital media resources, as data category, size of data, title and basic descriptor etc.Ontology library storage the ontology knowledge set of Digital Media.The mapping relations of semantic marker library storage material and Ontology.These two kinds of modes of original media material that are stored in strange land with integration that resource registering module is uploaded by registered user increase the media materials of digital media resource cloud.After source material registration, utilize labeling module under the guidance of Ontology, the original media material in media database is mapped to the concept in one or more ontology libraries.Resource retrieval module both can be retrieved media materials storehouse according to keyword, also can mark storehouse by query semantics media content is retrieved.

Consult the constructed Ontology view of Fig. 2 the present invention, this stereogram utilizes ontology services, retrieves the relation between all concepts and concept and concept in ontology library, by the form of tree structure, to user, represents result for retrieval.Body tree is comprised of Top-level Ontology and domain body: Top-level Ontology is described is the most general concept and the relation between concept, irrelevant with concrete application, and domain body is described is concept in specific area and the relation between concept.The Top-level Ontology of this Ontology tree draws after analyzing WordNet.Abstract concept from top layer of WordNet, things (thing) starts, by extracting the level (upper/the next relation) of implication relation in WordNet, formed an original body tree, the more original body tree of this class is repaired, translated just formed Top-level Ontology.On the basis of Top-level Ontology, by having researched and analysed the semantic feature of mass digital media resource, the related notion of some Digital Medias is expanded, form domain body.In body, each node is a concept class.Total node of semantic tree is " Thing(things) ", and other all classes are all the subclasses of root class " Thing(things) ".In concrete subclass, construct domain body tree.

Semantic tagger mainly contains label for labelling and object outline mark.Label for labelling is consulted shown in Fig. 3, system manager and user are under the constraint of ontology library, can add label for labelling to the various types of digital media resources in digital media resource cloud, these labels can be the nodes in semantic base, also can be constructed voluntarily by user.User can also submit the request that increases semantic node to, by keeper's audit by after can enrich Ontology storehouse.Object outline mark, only for image resource, is consulted shown in Fig. 4, and user utilizes semantic tagger instrument to depict the profile information of objects in images, then this object is added semantic.No matter be label for labelling or object outline mark, the digital media resource producing by mark and the mapping relations of semantic node are all stored in the semantic tagger table of database.The data set data set of the Digital Media researchs such as the details of the object outline in object outline mark are stored in xml file, and these profile informations not only comprise Object Semanteme, also can be used as object extraction, scene analysis is used.The present invention is provided with a kind of user's incentive system, digital media resource cloud system is taked integration system, system user is uploaded digital media resource and is added mark reward points to digital media resource, and integration can be used for downloading Digital Media related tool in resource in storehouse and use system.Motivational systems user shares Digital Media media resource and adds mark to resource in storehouse.Improve the degree of share of resource, creating knowledge resource (as academic data set) etc.

Consult Fig. 5 database structure schematic diagram of the present invention, black arrow represents the corresponding relation of the field of different list structures.Materials database, comprise the tables such as " t_picture(image) ", " t_vide(video) ", " t_model(model) ", the essential informations such as the ID of entity record digital media resource, URL address, title, digital media resource is when warehouse-in, the record of a resource of corresponding generation, the type stores of corresponding its resource is in these tables.Ontology storehouse, comprises tables such as " t_onology(semantic) ", for storing Ontology tree.Article one, a semantic node in body tree stored in record, recorded the information such as ID of the ID of semantic node, semantic title, father node.The tree structure of semantic table is determined according to internodal set membership.Semantic tagger storehouse, comprises tables such as " t_annotation(marks) ", records the mapping relations of digital media resource and semantic node.Each mapping relations is exactly a mark.The type of the mark object of this mark of resourceType field record in " t_annotation(mark) " table, in figure, dotted line represents, can be any one of Digital Media material Four types.The ID of resourceID field record mark object in corresponding material table in table.The record of the object of just can be in the storehouse unique calibration really of resourceType+resourceID note.OnologyID field in table, that deposit is the ID of body, in figure, dotted line represents, points in Ontology storehouse a body in " t_onology(is semantic) " table, namely the semanteme of this mark.

Consult the material retrieval flow figure of Fig. 6 Digital Media of the present invention.Digital media data Ku Zhong Digital Media material is supported the retrieval of two kinds of modes: simple retrieval (keyword retrieval) and semantic retrieval.When receiving that an inquiry request is, first the type of judgement request, if simple retrieval mates fields such as the title of raw data Ku Zhong Digital Media material table, descriptions with the keyword in inquiry request, retrieves result.If semantic retrieval, first with the search key coupling Ontology storehouse in inquiry request, obtain the semantic node of a coupling, recycle this semantic node retrieval semantic marker storehouse, obtain comprising these semantic all semantic tagger records, finally in raw data storehouse, inquiring each semantic tagger and record corresponding Digital Media material, has been both semantic retrieval result.Semantic retrieval is consulted shown in Fig. 7, at semantic browsing pages, clicks the node " competitive sports " in the Ontology tree of left side, so retrieved, comprises " competitive sports " this semantic material.

The not detailed disclosed part of the present invention belongs to the known technology of this area.

Although above the illustrative embodiment of the present invention is described; so that those skilled in the art understand the present invention; but should be clear; the invention is not restricted to the scope of embodiment; to those skilled in the art; as long as various variations appended claim limit and definite the spirit and scope of the present invention in, these variations are apparent, all utilize innovation and creation that the present invention conceives all at the row of protection.

Claims (5)

Translated fromChinese

1.一种云环境下数字媒体内容资源的集成共享方法，其特征在于以下步骤：1. a method for integrating and sharing digital media content resources under a cloud environment, characterized in that the following steps:步骤（1）、定义并实现数据库访问接口，利用云环境下异构数字媒体数据库访问与集成中间件，对不同地区的原始数字媒体素材进行统一的管理；Step (1), define and implement the database access interface, and use the heterogeneous digital media database access and integration middleware in the cloud environment to uniformly manage the original digital media materials in different regions;步骤（2）、提出数字媒体素材的本体概念分类及表示，以WordNet为本体蓝本，构建了一个针对不同应用领域的数字媒体素材语义本体库；Step (2), put forward the ontology concept classification and representation of digital media materials, and build a digital media material semantic ontology library for different application fields with WordNet as the ontology blueprint;步骤（3）、通过语义本体树节点与原始数字媒体素材建立对应关系，提供一种规范化语义描述方法，实现对原始数字媒体素材，在不同粒度下的语义标注。Step (3), establishing a corresponding relationship between the nodes of the semantic ontology tree and the original digital media material, providing a standardized semantic description method, and realizing the semantic annotation of the original digital media material at different granularities.2.如权利要求1所述的云环境下数字媒体内容资源的集成共享方法，其特征在于：（1）中的数字媒体数据库访问与集成中间件整合的包括图像资源、视频资源、三维模型资源、音频资源、数据集资源和软件资源在内的广域分布的异构的数字媒体资源。2. The method for integrating and sharing digital media content resources in a cloud environment according to claim 1, characterized in that: (1) the integration of digital media database access and integration middleware includes image resources, video resources, and 3D model resources Widely distributed heterogeneous digital media resources including audio resources, data set resources and software resources.3.如权利要求1所述的云环境下数字媒体内容资源的集成共享方法，其特征在于：（1）中定义并实现数据库访问接口是建立一种统一的mysql数据库格式，利用云环境下异构数字媒体数据库访问与集成中间件，在数字媒体数据库的原始素材子库保存异构的数字媒体数据库的副本，并利用中间件的接口与异构的数字媒体数据库保持同步更新，这样可以保证数据格式的一致性，避免过多的数据接口和数据转换，以提高查询效率。3. The method for integrating and sharing digital media content resources in a cloud environment as claimed in claim 1, characterized in that: (1) defining and realizing the database access interface is to establish a unified mysql database format, using different The digital media database access and integration middleware, save the copy of the heterogeneous digital media database in the original material sub-library of the digital media database, and use the interface of the middleware to keep synchronously updated with the heterogeneous digital media database, so that the data can be guaranteed The consistency of the format avoids too many data interfaces and data conversions to improve query efficiency.4.如权利要求1所述的云环境下数字媒体内容资源的集成共享方法，其特征在于：（2）中的数字媒体素材的通过提取wordnet中的蕴涵关系的层次，形成语义本体库的顶层本体，并通过研究分析海量数字媒体资源的语义特征丰富数字媒体相关节点，建立领域本体，最终本体信息是以树形结构存入语义本体库中，建立了一颗完备的数字媒体领域的本体语义树，本体具有面向数字媒体领域的通用性、完备性以及可扩展性。4. The method for integrating and sharing digital media content resources in a cloud environment as claimed in claim 1, characterized in that: the digital media materials in (2) form the top layer of the semantic ontology library by extracting the layers of implication relations in wordnet Ontology, and through research and analysis of the semantic features of massive digital media resources to enrich digital media-related nodes, establish a domain ontology, and finally store ontology information in a tree structure in the semantic ontology library, and establish a complete ontology semantics in the field of digital media The tree and ontology are universal, complete and extensible for the field of digital media.5.如权利要求1所述的云环境下数字媒体内容资源的集成共享方法，其特征在于：（3）中的不同粒度下的语义标注是指标签标注和对象轮廓标注，利用语义标注工具，对媒体素材中的对象、动作以及场景语义进行标注，从而建立起数字媒体资源到领域本体的映射关系，进而有效地按照本体语义去管理和检索数字媒体数据。5. The method for integrating and sharing digital media content resources in a cloud environment as claimed in claim 1, characterized in that: (3) semantic annotation at different granularities refers to label annotation and object outline annotation, using semantic annotation tools, Mark the objects, actions, and scene semantics in the media material, so as to establish the mapping relationship between digital media resources and domain ontology, and then effectively manage and retrieve digital media data according to ontology semantics.

Images