18 flash noflash

Upload: anand-sharma

Post on 03-Jun-2018

228 views

Category:

Documents


0 download

TRANSCRIPT

  • 8/12/2019 18 Flash Noflash

    1/18

    12/6/2013

    1

    3 Final Problems

    Photography in low light

    Photography in bad weather

    Photo forgeries

    Digital Photography with

    Flash and No-Flash Image Pairs

    G. Petschnigg, M. Agrawala, H. Hoppe, R. Szeliski, M.

    Cohen, K. Toyama

    E. Eisemann and F. Durand

    Photography in Low Light

    Using available

    ambient light:

    + natural lighting

    - high noise

    - color needs

    white balancing

    - blur

    No-flash

    Adding Lighting Shows Details

    Using flash:

    + details

    + color

    + low noise

    - flat/artificial

    - flash shadows

    - red eye

    Flash

  • 8/12/2019 18 Flash Noflash

    2/18

    12/6/2013

    2

    Flash + No-Flash

    Combine best aspects of each image

    Flash

    No-flash

    Result

    Flash + No-Flash Approach

    + original lighting+ details/sharpness+ noise removal+ color

    Result

    No-flash

    Use no-flash image to relight flash image

    (Petschnigg), or use flash image to relight

    no-flash image (Eisemann)

    Acquisition Process

    Lock Focus

    & Aperture

    1

    time

    Acquisition Process

    1/30 s

    ISO 3200

    No-Flash Image

    Large Sensor Gain/ISOLock Focus

    & Aperture

    21

    time

  • 8/12/2019 18 Flash Noflash

    3/18

  • 8/12/2019 18 Flash Noflash

    4/18

    12/6/2013

    4

    Recombination

    Intensity

    Result

    Color

    Flash

    ~* ~Result

    Recombination: Intensity * Color = Original

    shadow removal

    Results

    No-flash

    Flash

    Results

    Result

    No-flash

    Flash

    Poor Result

    Flash

    No-flash

  • 8/12/2019 18 Flash Noflash

    5/18

    12/6/2013

    5

    CVPR 2009

    Photographs in Bad Weather

    Haze Mist

    RainFog

    Images Courtesy : Steve and Carol Sheldon

    Low contrast

    Aerial Perspective

    akaAtmospheric

    Perspective

    Objects farther in thedistance appear less

    saturated (whiter) and less

    sharp (blurrier) than those

    nearby

    The more atmospheric

    particles between the

    viewer and a distant object,

    the more light is scattered

  • 8/12/2019 18 Flash Noflash

    6/18

    12/6/2013

    6

    Distant Objects are DesaturatedAerial Perspective

    Leonardo, Virgin and

    St. Anne, 1510

    Color Perspective

    Distant objects tend toward blue, near objects toward red

  • 8/12/2019 18 Flash Noflash

    7/18

    12/6/2013

    7

    Heuristic: Haze-free images have higher contrast than hazy images

    Medium transmission

    Most local patches in haze-free outdoor images that do not

    contain sky contain some pixels that have very low intensities

    in at least one color channel!

  • 8/12/2019 18 Flash Noflash

    8/18

    12/6/2013

    8

    Haze-free images have most pixels in dark

    channel near 0

  • 8/12/2019 18 Flash Noflash

    9/18

    12/6/2013

    9

    Daytime, outdoor

    landscapes or cityscapes

    from Flickr

  • 8/12/2019 18 Flash Noflash

    10/18

  • 8/12/2019 18 Flash Noflash

    11/18

  • 8/12/2019 18 Flash Noflash

    12/18

    12/6/2013

    12

    Results

    input recovered image

    Digital Image Forensics:

    Detecting Forgeries

    1. Detecting photorealistic graphics

    2. Detecting manipulated images

    12/6/2013

  • 8/12/2019 18 Flash Noflash

    13/18

    12/6/2013

    13

    CG vs. Real: Can You Do It?

    fakeorfoto.com

    I got 5 out of 12 right

    Chance = 6 out of 12

    Digital Image Forensics:

    Detecting Forgeries

    http://www.life.com/archive/realfake

    I got 6/10

    Chance = 5/10

    Detecting Forgery -- Why It Matters: Trust

    Iconic Portrait of Lincoln (1860)Lincolns head on John Calhouns body

    Examples collected by Hany Farid: http://www.cs.dartmouth.edu/farid/research/tampering.html

    1989 composite of Oprah and Ann-Margret (without eithers permission)

    12/6/2013

    http://area.autodesk.com/fakeorfoto/http://www.life.com/archive/realfakehttp://www.cs.dartmouth.edu/farid/research/tampering.htmlhttp://www.cs.dartmouth.edu/farid/research/tampering.htmlhttp://www.life.com/archive/realfakehttp://area.autodesk.com/fakeorfoto/
  • 8/12/2019 18 Flash Noflash

    14/18

    12/6/2013

    14

    2008

    1930s: Stalin had disgraced comrades airbrushed out of his

    pictures

    http://www.cs.dartmouth.edu/farid/research/digitaltampering/

    http://www.newseum.org/berlinwall/commissar_vanishes/index.htm

    2000: black students face inserted into UW magazine

    http://www.cs.dartmouth.edu/farid/research/digitaltampering/

    Pulitzer Prize winning photograph of Kent State killing (1970)

    12/6/2013

    http://www.cs.dartmouth.edu/farid/research/digitaltampering/http://www.newseum.org/berlinwall/commissar_vanishes/index.htmhttp://www.newseum.org/berlinwall/commissar_vanishes/index.htmhttp://www.cs.dartmouth.edu/farid/research/digitaltampering/http://www.cs.dartmouth.edu/farid/research/digitaltampering/http://www.newseum.org/berlinwall/commissar_vanishes/index.htmhttp://www.newseum.org/berlinwall/commissar_vanishes/index.htmhttp://www.cs.dartmouth.edu/farid/research/digitaltampering/
  • 8/12/2019 18 Flash Noflash

    15/18

    12/6/2013

    15

    2003: Long-time staff photographer for LA Times was fired for this one

    Published photo

    Kerry at Rally for Peace 1971 Fonda at rally in 1972

    Caption: Actress and Anti-war

    activist Jane Fonda speaks to a

    crowd of Vietnam veterans, as

    activist and former Vietnam vet John

    Kerry listens and prepares to speak

    next concerning the war in

    Vietnam. (AP Photo)

    Detecting Forgery: Cloning

    Exposing Digital Forgeries by Detecting Duplicated

    Image Regions

    A.C. Popescu and H. Farid

    Technical Report, TR2004-515, Dartmouth College, Computer

    Science

    Detecting Forgery: Retouching

    Exposing Digital Forgeries in Color Filter Array

    Interpolated Images

    A.C. Popescu and H. Farid

    IEEE Transactions on Signal Processing, 53(10):3948-3959,

    2005

    2005: Pres Bush scribbles a note to C. Rice

    during UN Security Council Meeting

    12/6/2013

  • 8/12/2019 18 Flash Noflash

    16/18

    12/6/2013

    16

    Demosaicing Prediction

    In demosaicing, RGB values are filled in based on

    surrounding measured values

    Filled in values will be correlated in a particular way for

    each camera Local tampering will destroy these correlations

    Farid: Photo Fakery

    and Forensics 2009

    Detecting Forgery: Lighting/Shadows

    Exposing Digital Forgeries by Detecting

    Inconsistencies in Lighting

    M.K. Johnson and H. Farid

    ACM Multimedia and Security Workshop, New York, NY, 2005

    Detecting Forgery: Lighting/Shadows

    Exposing Digital Forgeries by Detecting

    Inconsistencies in Lighting

    M.K. Johnson and H. Farid

    ACM Multimedia and Security Workshop, New York, NY, 2005

    Estimating Lighting Direction

    1 Method: 2D direction from occluding contour

    Provide at least 3 points on occluding contour (surface

    has 0 angle in Z direction)

    Estimate light direction from brightness

    Estimate

    Ground

    Truth

    12/6/2013

  • 8/12/2019 18 Flash Noflash

    17/18

    12/6/2013

    17

    Detecting Inconsistencies in Lighting

    Fake photo

    Real photo

    Detecting Forgery: Lighting/Shadows

    Lighting: Specular Highlights in the Eye

    M.K. Johnson and H. Farid, Exposing Digital Forgeries Through Specular

    Highlights on the Eye,2007

    Estimating Lighting from Eyes

    12/6/2013

    http://www.cs.dartmouth.edu/farid/publications/ih07.htmlhttp://www.cs.dartmouth.edu/farid/publications/ih07.htmlhttp://www.cs.dartmouth.edu/farid/publications/ih07.htmlhttp://www.cs.dartmouth.edu/farid/publications/ih07.html
  • 8/12/2019 18 Flash Noflash

    18/18

    12/6/2013

    18

    Summary

    Digital forgeries are an increasingly major

    problem as it becomes easier to fake images

    A variety of automatic and semi-automatic

    methods are available for detection of well-done

    forgeries

    Checking lighting consistency

    Checking demosaicking consistency (for high quality

    images)

    Checking JPEG compression level consistency (for

    low quality images)