Improving Image Matting using
Comprehensive Sampling Sets
CVPR2013 Oral
Outline




Introduction
Approach
Experiments
Conclusions
Introduction

Accurate extraction of a foreground object from
an image is known as alpha or digital matting.
Introduction

Applications
Introduction
Compositing Equation
Observed color of pixel z
Background color of pixel z
Foreground color of pixel z
Alpha value of pixel z
Introduction



Range of α : [ 0, 1]
α =1 , foreground.
α =0 , background.
Introduction


ill-posed problem
Typically, matting approaches rely on constraints
 Assumption on image statistics
 User constraints like Trimap
Unknown Region
Known Background
Known Foreground
Introduction

Current alpha matting approaches can be
categorized into
1.
alpha propagation based method
2.
color sampling based method
Introduction

Alpha propagation based method

Assume that neighboring pixels are correlated under
some image statistics and use their affinities to
propagate alpha values of known regions toward
unknown ones.
Introduction


Color sampling based method
 collect a set of known foreground and
background samples to estimate alpha values
of unknown pixels.
The quality of the extracted matte is highly
dependent on the selected samples.

missing true samples problem
Introduction
Approach





Gathering comprehensive sample set
Choosing candidate samples
Handling overlapping color distributions
Selection of best(F, B)pair
Pre and Post-processing
Approach


Gathering comprehensive sample set
For each region, a two-level hierarchical
clustering is applied.


first level, the samples are clustered with respect to
color
second level , respect to spatial index of pixels.
Approach

Gathering comprehensive sample set
Approach


Choosing candidate samples
Each pixel in the unknown region collects a set
of candidate samples that are in the form of a
foreground-background pair
Approach

Handling overlapping color distributions
Approach

Selection of best(F, B)pair
K : chromatic distortion
S : spatial statistics of the image
C : color statistics
Approach
Approach
Approach
Cohen's d
Approach

Pre-processing

An unknown pixel z is considered as foreground if,
for a pixel q ∈ F,
Trimap
Expanded Trimap
Approach

Post-processing

Eq. (2) is further refined to obtain a smooth matte by
considering correlation between neighboring pixels.
Cost function [5] consisting of the data term a and a
confidence value f together with a smoothness term consisting
of the matting Laplacian [10]

[10] A. Levin, D. Lischinski, and Y. Weiss. A closed-form solution to natural image matting. IEEE Transactions on
Pattern Analysis and Machine Intelligence, 30(1):228–242, 2007
[5] E. Gastal and M. Oliveira. Shared sampling for real time alpha matting. InProc. Eurographics ,
2010, volume 29, pages 575–584, 2010.
Experiments

www.alphamatting.com
Experiments

www.alphamatting.com
Experiments
Experiments
Experiments
Conclusions



A new sampling based image matting method
New sampling strategy to build a comprehensive set
of known samples.
This set includes highly correlated boundary samples
as well as samples inside the F and B regions to
capture all color variations and solve the problem of
missing true samples.