Compression de données 3D pour la transmission bas débit et

Report
UMR 5205
On the Efficiency of Image Metrics for
Evaluating the Visual Quality of 3D
Models
Guillaume Lavoué
Université de Lyon
LIRIS
Mohamed Chaker Larabi
Université de poitier
XLIM-SIC
Libor Vasa
University of
West Bohemia
An illustration
Original
Watermarking
Wang et al. 2011
Smoothing
Taubin, 2000
0.14
Watermarking
Cho et al. 2006
0.51
Simplification
Lindstrom, Turk
2000
0.40
Noise addition
0.62
Same Max Root Mean Square Error
(1.05 × 10-3)
0.84
Quality metrics for static meshes
Local curvature statistics
MSDM [Lavoué et al. 2006]
MSDM2 [Lavoué 2011]
[Torkhani et al. 2012]
Distorted
model
Local differences
of statistics
Matching
Original
model
Local Distortion Map
Spatial
pooling
Global Distortion Score
3
Our previous works
Distortion score
Why not using Image Quality Metrics?
Such image-based approach has been already used for driving
simplification
[Lindstrom, Turk, 2000][Qu, Meyer, 2008]
4
Our study
 Determine the best set of parameters to use for such imagebased quality assessment approach.
 Compare this approach to the most performing model-based
metrics.
5
Many parameters
 Which 2D metric to use?
 How many views, which views?
 How to combine the 2D scores?
 Which rendering, lighting?
 In our study, we consider:
o 6 image metrics
o 2 rendering algorithms
Around 100,000 images
o 9 lighting conditions
o 5 ways of combining image metric results
o 4 databases to evaluate the results
6
Image Quality Metrics
 Simple PSNR and Root Mean Square Error
 MSSIM (multi-scale SSIM) [Wang et al. 2003]
 VIF (visual information fidelity) [Sheikh and Bovik, 2006]
 IWSSIM (information content weighted SSIM) [Wang and LI, 2011]
 FSIM (feature similarity index) [Zhang et al. 2011]
State of the art algorithms
7
Generation of 2D views and
lightning conditions
42 cameras placed uniformly around the object
Rendering using a single white directional light
source
The light is either fixed with respect to the camera,
or with respect to the object
3 positions: front, top, top-right
 So we have 3*2 = 6 lighting conditions
 We also consider averages of object-light, camera-light and
global  9 conditions
8
Image Rendering Protocols
 We consider 2 ways of computing the normals, with or
without averaging on the neighborhood.
9
Pooling algorithms
 How to combine the per-image quality score into a single one?
 Minkowski norm is popular:
We also consider image importance weights
[Secord et al. 2011]
Perceptual model of viewpoint preference
 Surface visibility
10
The MOS databases
 The LIRIS/EPFL General-Purpose Database
 88 models (from 40K to 50K vertices) from 4 reference objects.
 Non uniform noise addition and smoothing.
The LIRIS Masking Database
 26 models (from 9K to 40K vertices) from 4 reference objects.
 Noise addition on smooth or rough regions.
The IEETA Simplification Database
 30 models (from 2K to 25K vertices) from 5 reference objects.
 Three simplification algorithms.
The UWB Compression database
 68 models from 5 reference objects
 Different kinds of artefacts from compression
11
Results and analysis
 Basically we have a full factorial experiments heavily used in
statistics to study the effect of different factors on a response
variable
 We consider 4 factors:
o The metric (6 possible values)
o The lighting (9 possible values)
o The pooling (5 possible values)
o The rendering (2 possible values).
 540 possible combinations
 We consider two response variables:
o Sperman correlation over all the objects
o Sperman correlation averaged per objects
12
Results and analysis
For a given factor associated with n possible values, we have n
sets of
paired spearman coefficients.
To estimate the effect of a given factor on the objective metric
performance, we conduct pairwise comparisons of each of its
value between the others (i.e. n(n-1)/2 comparisons).
We have paired values, so we can do better than a simple
comparison of the means.
 Statistical significance test (not Student but Wilcoxon signed rank test).
 We study the median of paired differences, as well as the 25th and 75th
percentiles.
13
Influence of the metrics
IWSSIM provides the best results
 FSIM and MSSIM are 2nd best, significantlky better than MSE
and PSNR.
 VIF provides instable results (see the percentiles).
14
Influence of the lighting
 Indirect illuminations provide better results
 Light has to be linked to the camera
 Object-front is not so bad, but not its performances are not
stable.
15
Influence of the pooling
Low values of P are better.
 Weights do not bring significant improvments.
16
Comparisons with 3D metrics
For easy scenarios: 2D metrics are excellent
However when the task becomes more difficult,
3D metrics are better
But, still, simple image-based metrics are better than simple
geometric ones.
17

similar documents