MM_tanvi_database Operations - Copy

download MM_tanvi_database Operations - Copy

of 16

Transcript of MM_tanvi_database Operations - Copy

  • 8/2/2019 MM_tanvi_database Operations - Copy

    1/16

    Presented By:

    Tanvi Sharma, 11-696

    M.E.(I.T), 2nd semester

    *

  • 8/2/2019 MM_tanvi_database Operations - Copy

    2/16

    *

    *MDBMS must offer all the above data types (i.e.

    raw data type , registering data type,

    descriptive data type)

    *Corresponding Operations for archival and

    retrieval

    *Mediarelated operations handled as a part of

    query languages (e.g. SQL)

  • 8/2/2019 MM_tanvi_database Operations - Copy

    3/16

    *Input (insert/record) : data will be written to thedatabase.

    Raw and registering data always required.

    Descriptive data can be attached later.

    MDBMS may face problems in choosing the properdisk/server(if during i/p operation,length of data not

    known)

    *Output (play) : reads the raw data from databaseaccording to the registering data.

    Huffman table required for decoding of a jpeg image

    Transmission of the raw data follows.

  • 8/2/2019 MM_tanvi_database Operations - Copy

    4/16

    *Modification : considers the raw data

    Should be done by an editor.

    For motion video : cutting with in/out fading.

    For audio data: volume, bass, treble, balance can alsobe modified.

    Modification function attributes are stored inregistering data.

    Data conversion from one format to another.(registering data must be modified along with raw

    data)

    Cntd..

  • 8/2/2019 MM_tanvi_database Operations - Copy

    5/16

    *

    Also understood as Transformation from one mediumto another medium (e.g. text-to-speech).

    The conversion function is implemented outside the

    DB.

    In editor:

    Reading data externally converting into another

    medium transformed data recorded back into thedatabase

  • 8/2/2019 MM_tanvi_database Operations - Copy

    6/16

    *Deletion: the consistency of the data must bepreserved. i.e. if a data type entry deleted, all related

    types should also be deleted

    *Comparison: many queries in MDBMS for search andretrieval of stored data.

    Individual patterns are simply comparedwith the storedraw data.

    (not very successful)

    With Pattern Recognition: Pattern from raw data may bestored as registering data and a comparison is based on thispattern. (only used for certain applications)

    Cntd

  • 8/2/2019 MM_tanvi_database Operations - Copy

    7/16

    *Comparisons based on content-oriented descriptive

    data.

    e.g. user enters a nominal phrase with a limited

    set of words. MDBMS converts this input into

    predicates.

    ( synonyms are used and are managed by system)

    *Evaluation: The goal is to evaluate the raw and

    registering data to generate the correspondingdescriptive data.

    e.g. during storage of facsimile documents, OCRcanbe used.

  • 8/2/2019 MM_tanvi_database Operations - Copy

    8/16

    *

    *Main issue for implementation and usage of an

    MDBMS is the Database Model

    *The data types and operations can be integrated

    in both; relational model and object-oriented

    model.

    *Abstract data types of all media can be defined

    with descriptive attributes according to their

    formats.

  • 8/2/2019 MM_tanvi_database Operations - Copy

    9/16

    *

    e.g.

    Attribute name Attribute Type Attribute Value

    Height: Integer; 480

    Width: Integer; 640

    Encoding: Uncompressed; YUV

    Stream_encoding: S_mode; PAL

    Pixel_depth: Integer; (8,8,8)

    Rate: Signed integer; 25

    Attributes of an uncompressed video

  • 8/2/2019 MM_tanvi_database Operations - Copy

    10/16

    *

    *Simplest possibility to implement a

    multimedia database. (since all attributes

    of different media are defined in

    advance)

    Main advantage: Compatibility withcurrent database applications

  • 8/2/2019 MM_tanvi_database Operations - Copy

    11/16

    Student (Admission_number Integer,

    Name String,

    Picture Image,

    Exercise_device_1 Video,

    Exercise_device_2 Video)

    Athletics (Admission_number Integer,

    Qualification Integer,

    high_jump Video,

    mile_run Video)

    Swimming (Admission_number Integer,crawl Video)

    Analysis (Qualification Integer,

    Error_Pattern String,

    Comment Audio)

    Attributes:picture,

    exercise

    devicesDifferent media types: image, video

    Other entries of database

    Analysis of different types of relational models

    *

  • 8/2/2019 MM_tanvi_database Operations - Copy

    12/16

    *

    *Type 1 Relational model: the value of a certainattribute can be fixed over the particular set of

    corresponding attribute types.

    e.g. frame rate of a motion video can be fixed

    *Type 2 Relational model : the variable number ofentries can be defined

    e.g. individual disciplines (swimming and athletics)

    of each admitted student are identified through their

    admission numbers

  • 8/2/2019 MM_tanvi_database Operations - Copy

    13/16

    *

    *Type 3 Relational model : in addition to fixed

    values of attributes and variable number ofentries, an entry can simultaneously belong to

    several relations.

    e.g. a video entry of a student performing ahigh jump can be assigned to the relation

    athletics for qualification purposes as well as

    to the relation analysis for an educationalapplication with analysis of typical errors.

    *

  • 8/2/2019 MM_tanvi_database Operations - Copy

    14/16

    *

    *Instead of defining relations, classes with objects are

    defined These objects are put into relations via a class hierarchy

    *Semantic specialization of classes and objects.

    e.g. Sports institute : main class

    Athletics, swimming: subclasses of sports institute

    Students : objects of each subclass

    *Offer good information navigation and flexible

    presentation*Important query operations (e.g. getting element from a

    set, include an element into a set) are incompletely

    supported(unlike relational model)

    *

  • 8/2/2019 MM_tanvi_database Operations - Copy

    15/16

    *

    *Jasmine is an Object-Oriented MMDB and has anapplication known as Studio as its development

    environment.

    *It comes with four multimedia classes:

    *Picture

    *Image

    *Video

    *Audio

    Manipulation and compression facilities .

    Has been made to fit well withJava Media Framework

  • 8/2/2019 MM_tanvi_database Operations - Copy

    16/16

    THANK YOU