Introduction MM Processing S08

download Introduction MM Processing S08

of 70

Transcript of Introduction MM Processing S08

  • 8/17/2019 Introduction MM Processing S08

    1/70

    CS 4763 Fundamentals of Multimedia Systems

    - Introduction to Image Processing

    Qi Tian

    Computer Science Department

    University of Texas at San Antonio

    [email protected]://www.cs.utsa.edu/~qitian/

  • 8/17/2019 Introduction MM Processing S08

    2/70

    Image Processing

    Manipulation of multidimensional signals− image (photo)

    − video

    − CT, MRI

    − Fluid flow

    ),(  y x f 

    ),,( t  y x f 

    ),,,( t  z y x f 

    ),,,( t  z y xv

  • 8/17/2019 Introduction MM Processing S08

    3/70

     A Typical Image Processing System

    object observe digitize store process Refresh

    /storeRecord 

    output

    DisplayImaging

    systems

    Sample and

    quantize

    Digital

    storage

    (disk)

    Digital

    computer 

    On-line

     buffer 

    X-ary, radar imaging, infrared

    imaging, ultrasound imaging,

    medical imaging, geophysicalimaging

    A/D

  • 8/17/2019 Introduction MM Processing S08

    4/70

    Fundamentals of Image Processing

    Representation – acquisition, digitization, and display to mathematical

    characterization of images for subsequent processing

     – a prerequisite for an efficient processing techniques such as

    enhancement, filtering, and restoration.

    Processing Techniques

     – Image compression, image restoration, and image reconstruction

     – Statistical image processing techniques

    Communications

  • 8/17/2019 Introduction MM Processing S08

    5/70

    Multimedia Processing Techniques

     – Coding/compressionStorage and communications JPEG, JPEG2000

    MPEG-1 (CD, mp3), MPEG-2 (HDTV, DVD)

    H.261, H.263

     – Enhancement, restoration, reconstruction feature extraction for image analysis and visual information display

    removal of degradation in an image, LS, ML, Max entropy, MAP

    2D -> 3D image MRI, CT, Radon transform

     – Analysis, detection, recognition, understandingquantitative measurements from an image to produce a description on it

     – Visualization

  • 8/17/2019 Introduction MM Processing S08

    6/70

     Advanced Processing Techniques

    Statistical processing techniques – Hidden Markov model (HMM)

     – Probabilistic graphical models

     – Bayesian networks

     – Markov random field 

    Many applications to speech recognition, pattern classification, data

    compression, and channel coding, etc.

  • 8/17/2019 Introduction MM Processing S08

    7/70

    History of Image/Video Coding

    1950

    1960

    1970

    1980

    1990

    2000+

    Math PR, CV, CG

    Fractal 3-D Model based

    coding

    Signal ProcessingBased

    PCM

    DPCM

    Transform Coding

    VQ

    Subband CodingWavelets

  • 8/17/2019 Introduction MM Processing S08

    8/70

    Reference:

     – F. Nebeker, Signal Processing: The Emergency of a Discipline,

    1948-1998

     – IEEE History Center, 1998

    Broadband TV (NTSC)

    500 × 500 × 8 × 3

    × 30 bits/sec

    100 Mb/sec (compression is necessary!)Modem: 56Kb/sec

    Picture Element

     – Pixel West coast people in USC – Pel East people in MIT

  • 8/17/2019 Introduction MM Processing S08

    9/70

    Image/Video Compression

    Signal-Processing Based:Encoder 

    ),(  y x f 

    H),(  y xg

    Signal

    Proc.

    Representation ),(  y xg

    Decoder  1H−),(ˆ  y x f ),(  y xg

  • 8/17/2019 Introduction MM Processing S08

    10/70

    Image/Video Compression

    3D Model-Based:Encoder 

    Representation  P

    Decoder 

    ),(  y x f 

    H

    Analysis Model

    Parameter  P

    Model

    ),(ˆ  y x f  P 3D

    Model

  • 8/17/2019 Introduction MM Processing S08

    11/70

    Image/Video Compression

    Fractal-Based:Encoder 

    Representation S

    Decoder 

    ),(  y x f System S

    ),(  y x f 

    Find S for which is an Attractor.),(  y x f 

    SAny

    signal),(ˆ  y x f 

    Iteration

  • 8/17/2019 Introduction MM Processing S08

    12/70

    Image/Video CompressionStandard 

    Facsimile: Fax Group 1, 2, 3, 4

    JBIG (Joint Bi-level Image Expert Group)

    Images: JPEG (http://www.jpeg.org/)

    JPEG2000

    Video: H.261, H.263 P  × 64 Kb/s (P =1 ~ 30)MPEG 1 1.2 Mb/s Video, CD, MP3

    MPEG 2 1.2 – 20 Mb/s, sports, HDTV, DVD

    MPEG 4 1 kb/s → 1Mb/s, very low speed video

    coding, MultimediaMPEG 7 Multimedia description, audio/video

    MPEG 21 Multimedia framework 

    Based on Wavelet

    Transform

  • 8/17/2019 Introduction MM Processing S08

    13/70

     A de facto image for the past three decade for its rich texture

    Lena

  • 8/17/2019 Introduction MM Processing S08

    14/70

  • 8/17/2019 Introduction MM Processing S08

    15/70

    What are Challenging Problems inMultimedia Processing?

    Multimedia Processing is taken in a broad sense,

    including:

    Image/Video compression, enhancement, restoration,

    reconstruction, analysis, recognition, understanding,

    visualization, and synthesis/animation.

  • 8/17/2019 Introduction MM Processing S08

    16/70

    Examples

    Face modeling, detection, and recognition

    Emotion recognition

    Gesture recognition

    Gender/age/ethnicity recognition

    Audio-visual speech recognition

    Image/video superresolution

    Image/video browsing, indexing, and retrieval

    Biometrics

  • 8/17/2019 Introduction MM Processing S08

    17/70

    Face Related Research

    Face modeling Face detection

    Face recognition

    Facial expression recognition

  • 8/17/2019 Introduction MM Processing S08

    18/70

    Generic Face Model

    Texture

    mapping

    Face model morphing

  • 8/17/2019 Introduction MM Processing S08

    19/70

    Generic Face Model

    The generic face model is generated from a MRI data set

  • 8/17/2019 Introduction MM Processing S08

    20/70

    Customize A Genetic Face on AnIndividual

    Polygon Mesh: 2240 Vertices + 3946 Triangles.Polygon Mesh: 2240 Vertices + 3946 Triangles.

     Non Non--Uniform Rational BUniform Rational B--Splines (NURBS): 63 control points.Splines (NURBS): 63 control points.

  • 8/17/2019 Introduction MM Processing S08

    21/70

    The iFACE system in a distributed collaborative environment. (a)

    Avatar in the head mounted display, (b) avatar in the desk screen of

    MIC3E, (c) avatar in the main screen of MIC3E

     Avatar – talking head

    University of Illinois at Urbana-Champaign

  • 8/17/2019 Introduction MM Processing S08

    22/70

    Text-Driven Face Animation

    “We strive to make the meter onanimation production, and are

    always looking for new technology

    that will enable faster, more

    appealing character creation,”

    said Joel Kransove, Digital Director of

     Nickelodeon. (Source: Digital

    Producer)

  • 8/17/2019 Introduction MM Processing S08

    23/70

    Speech-Driven Face Animation

    “Game characters have becomesynthetic actors and dialogue is anessential element of the effect wecreate. The quality of the lip-

    synching can make or break thesense of reality,”

    said Scott Cronce, vice president and CTO at

    Electronic Art (Source: Gamepro)

  • 8/17/2019 Introduction MM Processing S08

    24/70

     Video-Driven Face Animation

  • 8/17/2019 Introduction MM Processing S08

    25/70

    Emotion Recognition

  • 8/17/2019 Introduction MM Processing S08

    26/70

    Emotion Recognition

  • 8/17/2019 Introduction MM Processing S08

    27/70

    Emotion Recognition

  • 8/17/2019 Introduction MM Processing S08

    28/70

    Face Detection Techniques

  • 8/17/2019 Introduction MM Processing S08

    29/70

    Face Detection Techniques

  • 8/17/2019 Introduction MM Processing S08

    30/70

    Face Recognition: Why it is easy?

  • 8/17/2019 Introduction MM Processing S08

    31/70

    Face Recognition: Why it is hard?

  • 8/17/2019 Introduction MM Processing S08

    32/70

    Beauty Check

    What Are the Causes and Consequences ofHuman Facial Attractiveness?

    Babyfaceness

    Symmetry

    Social perception

    Universities of Regensburg, Germany

  • 8/17/2019 Introduction MM Processing S08

    33/70

    Which is more attractive?

    Universities of Regensburg, Germany

  • 8/17/2019 Introduction MM Processing S08

    34/70

    Babyfaceness Large head 

    Large curved forehead 

    Facial elements (eyes,

    nose, mouth) located

    relatively low

    Large, round eyes

    Small, short nose

    Round cheeks

    Small chinKate Moss4-year old girl

    Include mature female features: high, prominent cheek bones and

    concave cheeks

  • 8/17/2019 Introduction MM Processing S08

    35/70

    Which one is cuter?

  • 8/17/2019 Introduction MM Processing S08

    36/70

    Miss Germany (2002)

  • 8/17/2019 Introduction MM Processing S08

    37/70

    A selection of the 22 contestants of the final round of

    the contest

  • 8/17/2019 Introduction MM Processing S08

    38/70

  • 8/17/2019 Introduction MM Processing S08

    39/70

    Image Analysis

    Texture synthesis and transfer 

    Image Super-resolution

    Image Repairs

    Illumination/Lighting changes and transfer 

  • 8/17/2019 Introduction MM Processing S08

    40/70

    Texture Synthesis and Transfer

    +

    SIGGRAPH’01 Effros & Freeman, MIT, 2001

    synthesis

    transfer 

  • 8/17/2019 Introduction MM Processing S08

    41/70

    Texture Synthesis and Transfer 

  • 8/17/2019 Introduction MM Processing S08

    42/70

    Image Superresolution

    True Sub-sampled 

    Intelligent guess about details of texture

  • 8/17/2019 Introduction MM Processing S08

    43/70

    Image Superresolution

    Gaussian filter  Bicubic interpolation

  • 8/17/2019 Introduction MM Processing S08

    44/70

    Image Superresolution

    Median filter  Wiener filter 

  • 8/17/2019 Introduction MM Processing S08

    45/70

    Image Superresolution

    Dynamic resolutionenhancement

    Amos Storkey

    True

  • 8/17/2019 Introduction MM Processing S08

    46/70

    Image Repairs

  • 8/17/2019 Introduction MM Processing S08

    47/70

    Image Repairs

    Original Image

    Result

    Segmentation

    Image synthesis

     based on Tensor

    Voting

    Curve connection

  • 8/17/2019 Introduction MM Processing S08

    48/70

    Image Repairs

  • 8/17/2019 Introduction MM Processing S08

    49/70

    Illumination Effects on Images

  • 8/17/2019 Introduction MM Processing S08

    50/70

    Relighting – Basic Algorithm

    Step 2: Approximate radiance environment map

    Step 3: Synthesize novel appearance by adjustingthe 9 spherical harmonic coefficients

    Step 1: Align image with generic 3D face model

  • 8/17/2019 Introduction MM Processing S08

    51/70

    Lighting Transfer

    input target results

  • 8/17/2019 Introduction MM Processing S08

    52/70

    Image/Video Retrieval

    Image database

    CBIR b d l

  • 8/17/2019 Introduction MM Processing S08

    53/70

    CBIR based on color, texture,shape/structure

    MARS: Multimedia Analysis and Retrieval System

    metadata

    User

    Interface

    Similarity

    ranking

    memory

    Featureweighting

    Visual

    C++

    Feature

    Extraction

    C/C++  Color 

    Texture

    structure

  • 8/17/2019 Introduction MM Processing S08

    54/70

    State-of-the-art CBIR Systems

    QBIC (IBM), PhotoBook (Media Lab), Netra (UCSB),VisualSeek (Columbia), PicHunter (NEC-NJ), Amore (NEC-

    CA), EI Niňo (Praja), MARS (UIUC), Virage (Virage Inc.),

    CORE, PictoSeek, Piction, InfoScope …

    Research Communities

    Computer Vision, Image/Video Processing, Library and

    Information Science, Database and Management Systems

    Leading Journals & Standard

    PAMI, ACM Multimedia, IJCV, CVIU

    MPEG-7

  • 8/17/2019 Introduction MM Processing S08

    55/70

    MARS using global features

  • 8/17/2019 Introduction MM Processing S08

    56/70

    Biometrics

    Security Threats:

    We now live in a global society of increasing desperate and dangerous

     people whom we can no longer trust based on identification documents

    which may have been compromised.

    A challenging Pattern Recognition Problem

    Enabling technology to make our society safer,

    reduce fraud and offer user convenience.

  • 8/17/2019 Introduction MM Processing S08

    57/70

    Too many passwords to remember

  • 8/17/2019 Introduction MM Processing S08

    58/70

    Identification Problems

    Identity Theft: Identity

    thieves steal PIN (e.g., dateof birth) to open credit card

    account, withdraw money

    from accounts and take out

    loans

    3.3 million identity thefts in

    U.S. in 2002; 6.7 million

    victims of credit card fraud 

    Surrogate representations of identity such as password

    and ID cards no longer suffice

  • 8/17/2019 Introduction MM Processing S08

    59/70

    Biometrics

    Automatic recognition of people on their

    distinctive anatomical (e.g., face, fingerprint, iris,

    retina, hand geometry) and behavioral (e.g.,

    signature, gait) characteristics.

    Person identification is now an integral part of the

    infrastructure needed for diverse business sectors

    such as banking, border control, law

    enforcement…

  • 8/17/2019 Introduction MM Processing S08

    60/70

    Biometric Applications

  • 8/17/2019 Introduction MM Processing S08

    61/70

    Biometric Applications

    There are ~500 million border

    crossing/year (each way) in the US

  • 8/17/2019 Introduction MM Processing S08

    62/70

    Want to charge it?

  • 8/17/2019 Introduction MM Processing S08

    63/70

    Biometric Characteristics

  • 8/17/2019 Introduction MM Processing S08

    64/70

    Biometric Market Growth

    International Biometric Group

  • 8/17/2019 Introduction MM Processing S08

    65/70

     “State-of-the-art” Error Rate

    False accept rate

    (FAR):

    Proportion of

    imposters

    accepted 

    False reject rate

    (FRR):

    Proportions of

    genuine users

    rejected 

  • 8/17/2019 Introduction MM Processing S08

    66/70

    Multibiometrics

  • 8/17/2019 Introduction MM Processing S08

    67/70

    Soft Biometrics

  • 8/17/2019 Introduction MM Processing S08

    68/70

    Privacy Concerns

  • 8/17/2019 Introduction MM Processing S08

    69/70

    Tracking

  • 8/17/2019 Introduction MM Processing S08

    70/70