Belief Propagation with Directional Statistics for Solving the Shape-from-Shading Problem
Summary (2 min read)
1 Introduction
- A known or inferred reflectance function provides the relationship between irradiance and surface orientation.
- This constrained scenario has been tackled many times since[2–7, to cite a few], and will again be the focus of this work.
- More recent methods include Worthington and Hancock[5], which iterated between smoothing a normal map and correcting it to satisfy the reflectance information; Prados et al[6], which solved the problem with viscosity solutions; and Potetz[7] which used belief propagation.
- Belief propagation estimates the marginals of a multivariate probability distribution, often represented by a graphical model.
2 Formulation
- Using previously given assumptions, of Lambertian reflectance, constant known albedo, orthographic projection, an infinitely distant light source and no interreflection the irradiance at each pixel in the input image is given by Ix,y = A(̂l · n̂x,y) (1) where Ix,y is the irradiance provided by the input image.
- The normal map can be integrated to obtain a depth map, a step with which the authors are not concerned.
- By substituting the dot product with the cosine of the angle between the two vectors you get Ix,y A = cos θx,y (2) where θ is therefore the angle of a cone around l̂ which the normal is constrained to[5].
- This leaves one degree of freedom per pixel that is not constrained by the available information.
- The authors propose a new SfS algorithm using such distributions within a belief propagation framework.
3 Belief Propagation
- Such an equation can be represented by a graphical model where each variable is a node and nodes that interact via ψ functions are linked.
- Message passing then occurs within this model, with messages passed along the links between the nodes.
- The method uses belief propagation to obtain the maximum a posteriori estimate of a pairwise Markov random field where each node represents the orientation of the surface at a pixel in the image.
5 Method
- The authors construct a graphical model, specifically a pairwise Markov random field.
- For each node the authors have an irradiance value.
- The formulation presented so far will converge to a bi-modal distribution at each node, with the modes corresponding to the concave and convex interpretations.
- Each level’s messages are initialised with the previous, lower resolution, levels messages.
6 Message Passing
- Doing this directly is not tractable, so the authors propose a novel three step procedure to solve this problem: 1. Convert the FB8 distribution into a sum of Fisher distributions.
- All three steps involve approximation, in practise this proves not to be a problem.
- To derive a Fisher-Bingham distribution from the convolved sum of Fisher distributions the authors first need the rotational component of the Bingham distribution, which they calculate with principal component analysis.
- This is irrelevant as multiplicative constants have no effect.
7 Results & Analysis
- The authors compare the presented algorithm to two others, Lee & Kuo[4] and Worthington & Hancock[5], using both synthetic and real data.
- Figure 1 gives the four synthetic inputs used, figure 2 gives the results and ground truth for just one of the four inputs.
- For the Mozart 90◦ input their approach consistently exceeds Lee & Kuo but does not do so well at getting a high percentage of spot on estimates as Worthington & Hancock.
- The presented algorithm doing poorly as the light source moves away from [0, 0, 1]T can be put down to the bias introduced to handle the concave/convex ambiguity[12].
- Looking at figure 5 Worthington & Hancock is quantitatively ahead, but looking at the actual output it is more blob than face, though some features are recognisable.
8 Conclusion
- The authors have presented a new algorithm for solving the classical shape from shading algorithm, and demonstrated its competitiveness with previously published algorithms.
- The use of belief propagation with FB8 distributions is in itself new, and a method for the convolution of a FB8 distribution by a Fisher distribution has been devised.
- The algorithm does suffer a noticeable flaw in that overcoming the convex/concave problem biases the result, making the algorithm weak in the presence of oblique lighting.
- An alternative solution to the current bias is an obvious area for future research.
Did you find this useful? Give us your feedback
Citations
31 citations
Cites methods from "Belief Propagation with Directional..."
...The Bingham distribution is widely used in analyzing palaeomagnetic data (Kunze and Schaeben, 2004; Onstott, 1980), computer vision (Haines and Wilson, 2008), and directional statistics (Bingham, 1974) and recently in robotics (Gilitschenski et al., 2014, 2016; Glover et al., 2012; Kurz et al.,…...
[...]
...The Bingham distribution is widely used in analyzing palaeomagnetic data (Kunze and Schaeben, 2004; Onstott, 1980), computer vision (Haines and Wilson, 2008), and directional statistics (Bingham, 1974) and recently in robotics (Gilitschenski et al....
[...]
23 citations
Cites methods from "Belief Propagation with Directional..."
...The Bingham distribution is widely used to analyze paleomagnetic data [17], computer vision [12] and directional statistics [3] and recently in robotics [18, 9, 10, 8]....
[...]
16 citations
Cites methods from "Belief Propagation with Directional..."
...[6, 7] describe a method for integrating binocular stereo with SFS that also uses LBP....
[...]
12 citations
Cites background or methods from "Belief Propagation with Directional..."
...Our untrained system is comparable to [6] and [23] while the trained one is quantitatively superior for more relaxed angular errors....
[...]
...Table 1 provides a quantitative comparison of our system with [6] and [23] for the reconstruction of the Mozart surface....
[...]
...Even without learning the weighting parameters, the system presented so far compares favorably against [6] and [23] while for the more relaxed angular errors, the trained system is quantitatively superior....
[...]
7 citations
Cites background or methods from "Belief Propagation with Directional..."
...More recently, several authors have posed shape-from-shading in terms of pairwise Markov Random Fields [6, 7 ,8]....
[...]
...Ground truth in first column, remainder show: proposed algorithm, [4] and [ 7 ] respectively....
[...]
...Haines and Wilson [ 7 ] describe surface normal direction probabilistically in terms of a Fisher-Bingham distribution for each pixel....
[...]
...Coil database). In Fig. 1, we show a novel view of the surfaces recovered using the proposed algorithm, Worthington and Hancock [4] and Haines and Wilson [ 7 ]....
[...]
...Fig. 1. Surfaces recovered from the input image shown in the top left panel of Fig. 2. From left to right: ground truth, proposed algorithm, [4], [ 7 ]....
[...]
References
1,879 citations
"Belief Propagation with Directional..." refers background or methods in this paper
...Synthetic inputs, derived from the set used by Zhang et al[8]....
[...]
...Zhang et al.[8] surveyed the area in 1999, concluding that Lee and Kuo[4] was the then state of the art....
[...]
1,560 citations
889 citations
"Belief Propagation with Directional..." refers background in this paper
...This results in less message passes being required for overall convergence[13]....
[...]
695 citations
634 citations
"Belief Propagation with Directional..." refers background in this paper
...Horn[1] introduced this problem with the assumptions of Lambertian reflectance, orthographic projection, constant known albedo, a smooth surface, no surface inter-reflectance and a single infinitely distant light source in a known relation with the photo....
[...]
Related Papers (5)
Frequently Asked Questions (8)
Q2. What have the authors stated for future works in "Belief propagation with directional statistics for solving the shape-from-shading problem" ?
An alternative solution to the current bias is an obvious area for future research.
Q3. What is the reason why the presented algorithm does poorly?
The presented algorithm doing poorly as the light source moves away from [0, 0, 1]T can be put down to the bias introduced to handle the concave/convex ambiguity[12].
Q4. How long did the algorithm take to produce the head image?
For the head image the run time is over 12 hours for Lee & Kuo, 54 minutes for Worthington and Hancock and 9.5 minutes for the presented algorithm on a 2Ghz Athlon.
Q5. What is the method used to obtain the maximum a posteriori estimate of a pairwise?
The method uses belief propagation to obtain the maximum a posteriori estimate of a pairwise Markov random field where each node represents the orientation of the surface at a pixel in the image.
Q6. What is the definition of the distribution to be marginalised?
To define the distribution to be marginalised two sources are used: the irradiance information (Eq. 2) and a smoothness assumption.
Q7. What is the irradiance at each pixel in the input image?
Using previously given assumptions, of Lambertian reflectance, constant known albedo, orthographic projection, an infinitely distant light source and no interreflection the irradiance at each pixel in the input image is given byIx,y = A(̂l · n̂x,y) (1)where Ix,y is the irradiance provided by the input image.
Q8. How do the authors define a distribution to be marginalised?
Using equations 2 and 13 the authors can define a distributionΩ[2k Ix,y A l̂,−kl̂̂lT ] (17)In principle ψp(x̂p), from equation 16, can be set to this Bingham-Mardia distribution to complete the model to be marginalised.