computational photography and capture: time-lapse video analysis & editing

44
Computational Photography and Capture: Time-Lapse Video Analysis & Editing Gabriel Brostow & Tim Weyrich TAs: Clément Godard & Fabrizio Pece

Upload: palmer

Post on 31-Jan-2016

80 views

Category:

Documents


0 download

DESCRIPTION

Computational Photography and Capture: Time-Lapse Video Analysis & Editing. Gabriel Brostow & Tim Weyrich TAs: Clément Godard & Fabrizio Pece. Today’s schedule. Computational Time-Lapse Video Bennett & McMillan, Siggraph 2007 Factored Time-Lapse Video - PowerPoint PPT Presentation

TRANSCRIPT

Page 1: Computational Photography and Capture:  Time-Lapse Video Analysis & Editing

Computational Photography and Capture:

Time-Lapse VideoAnalysis & Editing

Gabriel Brostow & Tim WeyrichTAs: Clément Godard & Fabrizio Pece

Page 2: Computational Photography and Capture:  Time-Lapse Video Analysis & Editing

Today’s schedule

• Computational Time-Lapse Video– Bennett & McMillan, Siggraph 2007

• Factored Time-Lapse Video– Sunkavalli, Matusik, Pfister, Rusinkiewicz

Siggraph 2007

• Video Synopsis & Indexing, – Pritch, Rav-Acha, Peleg, ICCV 2007

Page 3: Computational Photography and Capture:  Time-Lapse Video Analysis & Editing

Time-lapse Photography

• Definition: when frames are captured at a lower rate than the rate at which they will ultimately be played back.

• Compare to– Slow-motion– Bullet-time

Page 4: Computational Photography and Capture:  Time-Lapse Video Analysis & Editing

Time-lapse Filmmaker’s Challenges

• Lighting changes (strobing)• High frequency motion (missed action)

– Saturation & exposure

• Camera motion: intentional or not• Triggering / data storage / camera safety

• Useful? Editable?

Page 5: Computational Photography and Capture:  Time-Lapse Video Analysis & Editing

Does the old definition of time-lapse still hold?

Page 6: Computational Photography and Capture:  Time-Lapse Video Analysis & Editing

(Quick Overview of)Computational Time-Lapse Video

Bennett & McMillanSiggraph 2007

(Project web page)(though link seems to have died, so see my archive here)

Page 7: Computational Photography and Capture:  Time-Lapse Video Analysis & Editing

Virtual Shutter

Page 8: Computational Photography and Capture:  Time-Lapse Video Analysis & Editing
Page 9: Computational Photography and Capture:  Time-Lapse Video Analysis & Editing

Sampling in 1D

• Uniform• Interpolate linearly+ piece-wise

vs.• Same # of samples, but where needed!• Optimum polygonal approximation of digitized

curves, Perez & Vidal, PRL’94– Use Dynamic Programming to find global min.

Page 10: Computational Photography and Capture:  Time-Lapse Video Analysis & Editing

Video

Page 11: Computational Photography and Capture:  Time-Lapse Video Analysis & Editing

Completely User-controllable:Min-Change, Min-Error, Median, etc

Page 12: Computational Photography and Capture:  Time-Lapse Video Analysis & Editing

Factored Time-Lapse Video

Sunkavalli, Matusik, Pfister, Rusinkiewicz SIGGRAPH 2007

Project web page

Page 13: Computational Photography and Capture:  Time-Lapse Video Analysis & Editing

Outdoors: More Than Just Color vs Illumination?

Original No Shadows

Given:Daytime, under clear-sky conditions

Wishlist

• Remove shadows• Render new shadows• Change albedo• Change global illumination

Wishlist

• Remove shadows• Render new shadows• Change albedo• Change global illumination

Page 14: Computational Photography and Capture:  Time-Lapse Video Analysis & Editing

Outdoors: More Than Just Color vs Illumination?

Original No Shadows

Lighting due to sun? sky? Surface Normals?

Page 15: Computational Photography and Capture:  Time-Lapse Video Analysis & Editing

Formulation

• size: width x height x time• F: xyt volume of frames over time

• Isky: Accumulated intensity due to sky-light

• Isun: Accumulated intensity due to sun-light

• Ssun: Binary; is a pixel out of shadow?

Page 16: Computational Photography and Capture:  Time-Lapse Video Analysis & Editing

Intuition

Page 17: Computational Photography and Capture:  Time-Lapse Video Analysis & Editing
Page 18: Computational Photography and Capture:  Time-Lapse Video Analysis & Editing

Separate Sky first

• Photoshop! – Just to pick “non-surfaces”

• When is sun’s contribution == 0?– Leaving just sky’s contribution

Page 19: Computational Photography and Capture:  Time-Lapse Video Analysis & Editing

• “Definitely” shadow: median of darkest 20%, then threshold at 1.5x

Page 20: Computational Photography and Capture:  Time-Lapse Video Analysis & Editing

Bilateral Filter “Definitely” Shadow

Page 21: Computational Photography and Capture:  Time-Lapse Video Analysis & Editing
Page 22: Computational Photography and Capture:  Time-Lapse Video Analysis & Editing

How Did Skylight Change?

• Isky: Accumulated intensity due to sky-light

• Wsky: Sky-light image

• Hsky: Sky-light basis curve (1D function!!)– Whole scene got brighter/darker together

Page 23: Computational Photography and Capture:  Time-Lapse Video Analysis & Editing

• Blue: skylight curve (same one)

Page 24: Computational Photography and Capture:  Time-Lapse Video Analysis & Editing

Factorization

• Decompose appearance into – per-pixel W – H curve

• Alternating Constrained Least Squares (ACLS):– H(t) is held fixed while W is

optimized using least squares, then vice versa

Page 25: Computational Photography and Capture:  Time-Lapse Video Analysis & Editing

ACLS: Alternating Constrained Least Squares

• Inverse shade trees for non-parametric material representation and editing,– Lawrence et al., Siggraph 2006, code online

Page 26: Computational Photography and Capture:  Time-Lapse Video Analysis & Editing
Page 27: Computational Photography and Capture:  Time-Lapse Video Analysis & Editing

Original and Isky

Page 28: Computational Photography and Capture:  Time-Lapse Video Analysis & Editing

So Far - Explained Sky

Page 29: Computational Photography and Capture:  Time-Lapse Video Analysis & Editing

Now Decompose Sun’s Contributions

• Isun: Accumulated intensity due to sun-light

• Wsun: Sun-light image, per-pixel weights

• Hsun: Sun-light basis curve• : Shift-map; per pixel offset in time

Page 30: Computational Photography and Capture:  Time-Lapse Video Analysis & Editing
Page 31: Computational Photography and Capture:  Time-Lapse Video Analysis & Editing

• Isun: Accumulated intensity due to sun-light

• Wsun: Sun-light image, per-pixel weights

• Hsun: Sun-light basis curve• : Shift-map; per pixel offset in time

Page 32: Computational Photography and Capture:  Time-Lapse Video Analysis & Editing

Total sunlight contribution

Sunlight image(like on moon)

Shift map Sunlight

Page 33: Computational Photography and Capture:  Time-Lapse Video Analysis & Editing

Factorize Again!

• For each image, subtract off Isky(t)

• Ssun: Sky-light mask, serves as C, confidence map

• ACLS again– But add 3rd update phase, searching for

that minimizes reconstruction errorof scaled + offset H(t) vs. F(t)

Page 34: Computational Photography and Capture:  Time-Lapse Video Analysis & Editing

Green: sunlight curve (same one, but time-shifted)

Page 35: Computational Photography and Capture:  Time-Lapse Video Analysis & Editing

Isun(t)

Wsun

(time offset)

Page 36: Computational Photography and Capture:  Time-Lapse Video Analysis & Editing

Reconstruction

Page 37: Computational Photography and Capture:  Time-Lapse Video Analysis & Editing

Does It Work?

(black curves are calculated sunlight basis curves)

Page 38: Computational Photography and Capture:  Time-Lapse Video Analysis & Editing

Video: FactoredTimeLapseV

Page 39: Computational Photography and Capture:  Time-Lapse Video Analysis & Editing

What have we gained?

• Can– Remove shadows– Change albedo– Change amount of global illumination

• Compression• Render new shadows:

– We have 1 component of the per-pixel Normals, N

Page 40: Computational Photography and Capture:  Time-Lapse Video Analysis & Editing

Estimated Normals (1D)

Page 41: Computational Photography and Capture:  Time-Lapse Video Analysis & Editing

Manipulating Normals

Page 42: Computational Photography and Capture:  Time-Lapse Video Analysis & Editing

(View Results of)Video Synopsis & Indexing

Pritch, Rav-Acha, Peleg ICCV 2007, PAMI 2008

(Project web page)

Page 43: Computational Photography and Capture:  Time-Lapse Video Analysis & Editing

What to do with blob-tracks?

(Project web page)

Page 44: Computational Photography and Capture:  Time-Lapse Video Analysis & Editing

HDR Time-Lapse

• Essentially started in 2006 (manually at first)• More are popping up online