EE 396: Lecture 3 Ganesh Sundaramoorthi Feb. 15, 2011

0.1

Review

Physical World (X1 , Y1 , Z1 ) ∈ R3 R(X1 , Y1 , Z1 )

(X1 , Y1 ) ∈ Ω u(X1 , Y1 ) := R(X1 , Y1 , Z1 ) u(X2 , Y2 )

R(X2 , Y2 , Z2 )

I =u+η

u

u(X3 , Y3 )

R(X3 , Y3 , Z3 )

+

p(u) � ∝� � exp − 2σ1 2 Ω |∇u(x)|2 dx p

η(x) ∼ N (0, σ) iid x ∈ Ω

Object(s)

Imaging Plane Ω Affine Camera

Figure 1: Model of image formation for constructing our denoising algorithm. Figure 1 shows our model for image formation. Note that in general that the denoising problem is an ill-posed problem and cannot be solved without knowledge of how the image is formed, how the noise is generated, and a characterization of the possible true images. We have a simplified model of the image formation process where we have assumed that • visible points on the objects project to the imaging plane by the simple projection (X, Y, Z) → (X, Y ) ∈ Ω. This is true when the object(s) are very far from the imaging plane, and this a special case of what is known as an affine camera. For more information on the image formation process (which we will see again in more detail when we study image registration, see [2]). • The irradiance R, that is, the light incident on the surface is directly recorded on the imaging plane : u(X, Y ) = R(X, Y, Z). 1

• The error in the CCD device or any other noise/nuisances in the physical world are modeled by an additive noise on u, and the image we observe is I. • The noise η is modeled by a Gaussian iid random variables. • The prior probability on the space of possible true images u is proportional to the exponential of the norm gradient squared of u. Our goal is to recover u from I using the model above. We note that to design better denoising algorithms, we must improve the model above! Using the above assumptions, the Bayesian approach, and MAP estimation, we derived the following energy to be minimized : Z Z 2 (1) E(u) = − log p(u|I) = (I(x) − u(x)) dx + α |∇u(x)|2 dx, Ω



where E : U → R and

Z

U = {u ∈ C (Ω, R) :

|∇u(x)|2 dx < ∞,

2



∂u (x) = 0 for x ∈ ∂Ω}. ∂n

We computed the conditions for critical points of the following generalized functional: Z E(u) = F (u(x), ∇u(x), x) dx

(2)

(3)



and these are called the Euler-Lagrange equations : ( Fw (u(x), ∇u(x), x) − div (Fy (u(x), ∇u(x), x)) = 0 for x ∈ int(Ω) . ∂u for x ∈ ∂Ω ∂n (x) = 0

(4)

Note that these conditions are necessary for a local min/max of E, but are not sufficient. Next, we introduce conditions on F that will tell us when the solution to the Euler-Lagrange equations is a global minimizer of E.

0.2

Convexity

We introduce a very important property of functionals that guarantees that the functional does not have local minima that are not global minima. In addition, having the property of convexity leads to many efficient algorithms for minimizing the functional. Definition 1. A class of functionals F is a convex space if it is a subset of a vector space and for all t ∈ [0, 1] and u1 , u2 ∈ F, we have that (1 − t)u1 + tu2 ∈ F. Remark 1. Note that U defined in (2) is a convex space. Let’s verify. Let u1 , u2 ∈ U. Note that the second partials of u1 and u2 exist and are continuous by definition, and therefore the function (1 − t)u1 + tu2 has second partials and they are continuous. Also, Z |∇((1 − t)u1 (x) + tu2 (x))|2 dx = Ω Z Z Z 2 2 2 2 (1 − t) |∇u1 (x)| dx + t |∇u2 (x)| dx + 2t(1 − t) ∇u1 (x) · ∇u2 (x) dx. (5) Ω





2

Recall the inequality ab ≤ (a2 + b2 )/2 for a, b ∈ R (which is derived from (a − b)2 ≥ 0). Applying this, we find Z Z Z

Therefore, Z |∇((1 − t)u1 (x) + tu2 (x))|2 dx ≤ Ω

Z

Z

|∇u2 (x)|2 dx < ∞. (7)

|∇u1 (x)| dx + (t + t(1 − t))

((1 − t) + t(1 − t))

2

2

(6)







|∇u2 (x)|2 dx.

|∇u1 (x)|2 dx +

∇u1 (x) · ∇u2 (x) dx ≤

2

2





Therefore, ∇((1 − t)u1 (x) + tu2 (x) ∈ U, and so U is a convex space. We are now ready to define a convex functional. Definition 2. A functional E : U → R is a convex functional if U is a convex space, and if for all t ∈ [0, 1] and u1 , u2 ∈ U we have that E((1 − t)u1 + tu2 ) ≤ (1 − t)E(u1 ) + tE(u2 ).

(8)

It turns out that testing for convexity of E can be done by just testing convexity of F when E is in the form of (3). Proposition 1. The functional E defined in (3) is convex when F is convex with respect to the first two arguments, i.e., F (w, y, z) is convex in (w, y) for all z. Proof. Convexity of F in (w, y) means that for all x F ((1 − t)u1 (x) + tu2 (x), ∇((1 − t)u1 (x) + tu2 (x)), x) ≤ (1 − t)F (u1 (x), ∇u1 (x), x) + tF (u2 (x), ∇u2 (x), x). (9) Integrating the previous inequality yields that E is convex. There are equivalent notions of convexity of F when F satisfies additional assumptions of smoothness: Proposition 2. Let f : Rn → R be a convex function, and f ∈ C 2 (Rn , R) then for x, y ∈ Rn , the previous conditions are equivalent to 1. f (x) − f (y) ≥ ∇f (x) · (x − y) 2. let Hf (x) be the Hessian matrix at x, i.e., (Hf (x))ij = (fxi xj (x)), then Hf (x) is positive semidefinite, that is, for all y, y T Hf (x)y ≥ 0 where T denotes transpose. The above is typically convenient to check whether F is convex, but it only works for smooth functions! We will see an example of a non-smooth F in a few lectures. Let us now check whether F defined for our original denoising model F (w, y, z) = (I(z) − w)2 + α|y|2 .

(10)

leads to a convex functional E. Since F is smooth, we compute the second partials with respect to w, y: Fw = 2(w − I(z)), Fww = 2, Fy = αy, Fyy = αId, Fyw = 0 3

(11)

where Id is the 2 × 2 identity matrix. Therefore,  H(w,y) F (w, y, z) =

2 0 0 αId

 ,

(12)

which is postive semi-definite since α > 0. Moreover, the Hessian is strictly positive definite that is (w0 , y0 )T HF (w, y, z)(w0 , y0 ) > 0 for (w0 , y0 ) 6= 0. So E defined for our denoising model is strongly convex. Convexity implies that E has only global minima if there are minima at all. Indeed, we have the following theorem: Proposition 3. Let E : U → R be convex, and suppose that there exists u ∈ U such that dE(u) · v = 0 for all v ∈ V. Then u is a global minimizer of E. Proof. Let w ∈ U, and note that w − u ∈ V. By convexity, we have E(u + t(w − u)) ≤ (1 − t)E(u) + tE(w);

(13)

rearranging we have that

E(u + t(w − u)) − E(u) ≤ E(w) − E(u), t and taking the limit as t → 0, we find dE(u) · (w − u) ≤ E(w) − E(u),

(14)

(15)

but the left hand side is zero, and so E(u) ≤ E(w), and this is true for any w. Therefore, u is a global minimizer. Now back to our denoising problem, we know that E is convex and we have found the equation ( α∆u(x) = u(x) − I(x) x ∈ int(Ω) . ∂u (x) = 0 x ∈ ∂Ω ∂n

(16)

that must be solved in order to find a critical point of E. Now by the previous theorem, if a critical point exists, then it must be a global minimizer. Thus, we simply need to be able to show that (16) has a solution, and then automatically it will be a global minimizer. We will do this next lecture. For more details on convexity of see [1], and for convexity of functionals see [3].

References [1] Stephen Boyd and Lieven Vandenberghe. Convex Optimization. Cambridge University Press, 2004. [2] Y. Ma, S. Soatto, J. Kosecka, and S. Sastry. An invitation to 3-D vision, volume 6. Springer, 2004. [3] John L. Troutman. Variational Calculus and Optimal Control: Optimization with Elementary Convexity. Springer, 1996.

4

EE 396: Lecture 3 - UCLA Vision Lab

Feb 15, 2011 - The irradiance R, that is, the light incident on the surface is directly recorded ... partials of u1 and u2 exist and are continuous by definition, and ...

384KB Sizes 2 Downloads 324 Views

Recommend Documents

EE 396: Lecture 3 - UCLA Vision Lab
Feb 15, 2011 - (which we will see again in more detail when we study image registration, see [2]). • The irradiance R, that is, the light incident on the surface is ...

EE 396: Lecture 10-11
From your physics class, we know that the speed of the curve squared divided by the radius of curvature is the normal component of acceleration : cpp(p) · N(p) = |cp(p)|2. R(p). = |cp(p)|2κ(p). (20) where κ(p) is one over the radius of curvature;

EE 396: Lecture 2
Feb 12, 2011 - where I : Ω → R is the observed or measured image, and α > 0. The energy E is a functional that is defined on the class of functions U, which ...

EE 396: Lecture 4
where Id : U→U is the identity map, and so if u1,u2 ∈ U, then .... Recall from signal processing, that if we have a linear system that is also shift-invariant (some-.

EE 396: Lecture 5
Feb 22, 2011 - In an inner product space, we automatically get for free the Cauchy-Schwartz .... smoothing with and closeness to the original image I. 3The is a ...

EE 396: Lecture 8
Mar 8, 2011 - many fields including electrical engineering and financial data .... used symmetry of w, Fubini's Theorem to interchange order of integration, and ...

EE 396: Lecture 14-15
Calculating the posterior probability, p({Ri,ui}N i=1|I) using Bayes' Rule, and then calculating the MAP estimate is equivalent to minimizing the energy. E({Ri,ui}N.

EE 396: Lecture 9
Mar 12, 2011 - p(ui) (stochastic components of ui, αi and ηi, are independent from each other ) ... we could minimize in u, and a MAP estimate for u, which would.

EE 396: Lecture 13
Mar 29, 2011 - We shall now derive an algorithm whereby we can compute dS(x) for .... Lagrange equations are. ∇Lφ(γ) = ∇φ(γ(s)) − d ds. (φ(γ(s))γs(s)) = 0.

EE 396: Lecture 12
Mar 26, 2011 - Thus, along lines that are parallel to v the function remains ... x ∈ R. We simply follow a line parallel to v starting from (t, x) and follow the line ...

Lecture 3
Oct 11, 2016 - request to the time the data is available at the ... If you want to fight big fires, you want high ... On the above architecture, consider the problem.

Oxford Handbooks Online - Bjork Learning and Forgetting Lab - UCLA ...
As an alternative view, Koriat ..... It is our view that both types of accuracy should be evaluated whenever ...... Applied Cognitive Psychology, 27, 360-. 372.

Development of Three-Dimensional Object ... - UCLA Baby Lab
information in the visual array, and emerging manual and other ... as a unified object rather than two disjoint segments. (Kellman ... 2007). Infants at 2 and 4 months will complete a 3D ... month-olds (M age 5 118.4 days, SD 5 12.8) with 11.

Systems in Development: Motor Skill Acquisition ... - UCLA Baby Lab
habituated to a limited-view object and tested with volumetrically complete and incomplete .... only the sides seen in the limited view) that now rotated a full 360°.

Statistical learning across development: flexible yet ... - UCLA Baby Lab
Jan 11, 2013 - the developmental time course of visual statistical learning during ... of different statistical features allow varying degrees of associa- ...... Davis (Cambridge, MA: MIT Press), ... December 2012; published online: 11 Jan-.

Statistical learning across development: flexible yet ... - UCLA Baby Lab
Jan 11, 2013 - 1 Department of Psychology, University of California, Los Angeles, CA, USA. 2 Department of ... Sciences, Germany. Reviewed by: ... seemingly effortlessly, in their first years after birth? For exam- .... on a computer screen. For half

Oxford Handbooks Online - Bjork Learning and Forgetting Lab - UCLA ...
Keywords: metamemory, education, memory, learning, students ...... With the rising use of technology in the classroom, studies involving learners' perceptions of ...

Visual Recognition - Vision & Perception Neuroscience Lab - Stanford ...
memory: An alternative solution to classic problems in perception and recognition. ... of faces, objects and scenes: Analytic and holistic processes (pp. 269–294).

Lab 3: Structure - GitHub
Structure Harvester is very easy to use, and is all web-based! You simply upload your zip file and then click “Harvest!” It may take a few minutes to run.

EE - Lecture 5b - RoR - multiple projects.pptx
Must use LCM for each pair-wise comparison. • Procedure for ... If i* < MARR: eliminate project and go the next project. •. If i* ≥ MARR: go ... Rules: – No phones.

Lab 3 Example - GitHub
Download “polygonum.stru”'. • Look at “polygonum.stru” using a text editor. – Column 1 refers to individual ID (516 total individuals). – Column 2 refers to ...

Lecture 3.pdf
Page 1 of 36. Memory. It is generally agreed that there are three types of. memory or memory function: sensory buffers, short-term. memory or working memory, ...

EE 305 Electronic Devices and Design Lab
Phone: (256) 824-6469. Office: 410 Optics Building. Office Hours: Monday 11:00 AM - 12:00 AM, Tuesday 2:00 PM- 4:00 PM. Meeting .... Questions and Review: 1. ... Use equation 2.2 to plot the Bode diagram of the system. .... d) You are asked to design

Lab 3.pdf
A possible source of variation could be inaccurate analyzing of data,. but a change ... Yes because if the cuff is released too quickly, blood may rush into the arm.