3 4 Approxim of The RTE 79

3. Tissue Optics! 3.4 Approximations of the RTE! 3. Tissue Optics! 3.4 Approximations of the RTE!
3. Tissue Optics
3.4 Approximations of the RTE!
3.1 Basics in EM Optics
3.2 Tissue optical parameters
3.3 The radiative transport equation (RTE)
3.4 Approximations of the RTE
3.4.1 The Diffusion approximation!
3.5 Measurement methods
Georges Wagnières, ISIC, EPFL! Photomedicine, spring 2012! 1! Georges Wagnières, ISIC, EPFL! Photomedicine, spring 2012! 2!
Radiative Transport Equation! Diffusion Approximation!

Observation
Incident direction
Radiance L (W.m-2.sr-1) direction
ˆs!
sˆ
Loss:
Absorption +
scattering • For a dense medium containing mostly scatterers, the
Change of radiance L r
Elementary
diffusing element
transport equation can be simplified to the Diffusion
Equation!
• Can be solved analytically for special cases or more

generally by numerical techniques!
Gains: scattering + source • Validity of diffusion equation is limited to tissue cases

Integro-differential equation difficult to solve where the light has been highly scattered ( ). "
!Use the diffusion approximation equation. should be at least 10 times greater than!
Diffusion approximation! Radiant Energy Fluence Rate " (r,t)!

The integral of the radiance over all directions, called the
fluence rate "(r) [W/m2], has more practical significance than
The diffusion approximation can be solved analytically for the radiance itself.!
simple geometries. !
The diffusion approximation has been used extensively in

photon propagation modeling in tissue in the time, as well as in
the steady state domains.!
The fluence rate is defined as the radiant power incident on a
small sphere, divided by the cross sectional area of that
sphere!
Diffusion approximation! Solutions of the Diffusion Approximation"

(Selected examples)!
Simplified form of RTE at “diffusion limit”!
Assuming that L(r,s,t) can be approximated by:! • Interstitial "point" source in an infinite volume of tissue!
! !where J is the photon current!

• "Point" source at the air-tissue interface of a semi-infinite
For more details:"
A. J. Welch & M.J.C. van Gemert, Plenum, 1995!
volume of tissue!
– The source must be much beyond 1/µs’ (the light MUST be diffuse)!
• ~10/µs’ (~ 1 mm in biological tissue)! • Collimated "Broad" illumination perpendicular to the air-
• far from sources & boundaries! tissue interface of a semi-infinite volume of tissue!
• assume tissue is “macroscopically homogeneous”!
• Interstitial illumination with a cylindrical light distributor!
Interstitial "point" source in an infinite

volume of tissue!
Point source solution!
The solution to the diffusion equation for an
homogenous slab with a short pulse isotropic point
source S(r,t)=#(0,0) is:!
0.1 mm-1!
1 mm-1! "3 / 2 r2
# (r, t ) = c(4!Dct ) exp(" " µ a ct )
4 Dct
This Green’s function can be used to solve problems

' for a slightly more complex geometry!
Steady State Diffusion! Solution to stady state diffusion!

In steady state diffusion the time dependance is equal to
zero and the equation for a point source is:!
2
# 2 ! (r ) " µ eff ! (r ) = S (r )
This yields a solution for an infinite homogenous

medium as: !
2
1 Pµ eff 1
# (r ) = # (r = 0) exp(" µ eff ! r ) = exp(" µ eff ! r )
r 4$µ
' a r '
Point source solution" The Time-independent Fluence Rate!

Tissue / water interface!
If we consider a pencil beam incident perpendicular on This solution would yield the following fluence
a semi-infinite volume of tissue !
inside a semi-infinite homogenous medium:!
n1 #$ Incident light of power: P

n2
n1 = n2 '
Collimated “Broad” illumination Collimated “broad” illumination

perpendicular to the air-tissue interface" perpendicular to the air-tissue interface"
Semi-infinite volume of tissue! Semi-infinite volume of tissue (n=1.37)!
Model and Monte-Carlo simulation (S.L Jacques)!
Irradiance: "0 Diffusion

approximation
is not valid
n1= 1 anymore!
": Fluence rate Backscattering
to the interface!
n1= 1.37
" = "0ke -!effz
z '
Interstitial illumination with an infinitely long"

Interstitial illumination with a cylindrical light distributor!
cylindrical light distributor!
Jc [W/cm]: Output power per unit length of fiber!
r [cm]: Radial distance from the fiber axis!
ac [cm]: Fiber radius !
2 J c / !ac
"(r ) = K 0 (rµ eff )
K 0 (ac µ eff ) + 2 Dµ eff K1 (ac µ eff )
K0 and K1 are modified Bessel functions of the second kind!

(see next slide)!
'
Modified Bessel Functions of the Second Kind" Lines of constant fluence rate about a linear
(http://mathworld.wolfram.com/ModifiedBesselFunctionoftheSecondKind.html)! light applicator in a rat rhabdomyosarcoma!
Fluence rate
x!
Bessel functions ! fast decrease!
Source: J. Marijnissen!
3.4.2 The Kubelka-Munk Theory! Radiance!

The Kubelka-Munk theory describes the propagation of a It is the radiant power
uniform, diffuse radiance through a one-dimensional isotropic
(W) per unit solid angle
slab with no reflection at the boundaries.!
(steradian) about unit
F+(0)! F-(0)! vector, s and per unit area
z=0!
(m2) perpendicular to s.!
F+! F-+dF-!
z!
z+dz!
F++dF+! F-!
z=d!
F+(d)!
P. Kubelka, "New contributions to the optics of intensely light-scattering materials. Part I", J Opt Soc Amer,38,pp. 448-457
(1948)! It is diffuse when L(r, s) is the same in all directions s.!
P. Kubelka and F. Munk, "Ein Beitrag zur Optik der Farbanstriche", Zeitschr Techn Phys 12, pp. 593-601 (1931) !
Kubelka-Munk Theory! The Kubelka-Munk Theory!
The KM theory is a special case of the "many-flux" theory.! F! (0 ) The radiance in either direction encounters !
R=
It is based on a two flux model where F+ (0 ) • 2 losses due to absoption and scattering !
the radiation is composed of two • 1 gain due to scattering of photons from
F+(0)! oppositely directed diffuse radiation " F+(0)! the opposite direction!
F-(0)! F-(0)!
z=0! fluxes F+ and F-.! 2 z=0!
[F] = W/cm ! d F+ (z )
F+! F-+dF-! F+! F-+dF-! = !(K + S )F+ + S F!
z! z! dz
z+dz! Application criteria:! z+dz!
F++dF+! F-! • diffuse irradiance! F++dF+! F-! d F! (z )
• one-dimensional isotropic medium! = (K + S )F! ! S F+
z=d! z=d! dz
F+(d)! • no reflection at boundaries! F+(d)!
F+ (d ) with K and S the Kubelka-Munk
T=
Approximation: any collimated flux Fc % 0 after very short distance! F+ (0 ) absorption and scattering coefficients!
Kubelka-Munk Theory"
Kubelka-Munk Theory"
Special Solution! Reflection and Transmission!
Special solutions for a tissue with: !
1) a known slab thickness d !2) matched refraction indices ! The Kubelka-Munk absorption and scattering coefficients may
3) isotropic scattering !4) diffuse irradiance! be directly expressed in terms of the measured reflection R and
a ! sinh[b ! S ! (d " z )]+ b ! cosh[b ! S ! (d " z )] the transmission T.!
F+ (z ) = F(0 )
a ! sinh[b ! S ! d ]+ b cosh[b ! S ! d ] F! (0 ) F+ (d )
From! R = and! T= One can deduce!
F+ (0 ) F+ (0 )
sinh[b ! S ! (d " z )]
F" (z ) = F (0 )
a ! sinh[b ! S ! d ]+ b ! cosh[b ! S ! d ]
S=
1
( ln $ (
&1 ' R a ' a 2 ' 1 )#! and! K = (a " 1)! S
The total fluence rate is:! a 2 ' 1 ( d $% T !
"
"(z) = 2 # (F+ (z) + F$ (z))
K + S 1 + R 2 ! T2
(K + S ) with! a = =
with F(0) the incident flux, ! a =
S
and! b= (a 2
)
!1 S 2R
!
Measurement of the Reflection and Transmission! KM theory for collimated flux!

Two Flux KM Model! Clinical Applications!
• uniform, diffuse irradiance ! • at least 3 fluxes: two diffuse
• exclusion of collimated flux! fluxes and one collimated flux
(e.g. from laser irradiation)!
• no reflection at boundaries! • specular as well as internal
diffuse reflection at the air-tissue
(tissue-air) boundary!
• one-dimensional isotropic • two or more layers of tissue!
medium!
Three flux theory!
Three flux Approach! The three flux approach!

Collimated flux Fc, and diffuse The matrix-based presentation can be used to express the fluxes at
fluxes F+ and F-:! z=d in terms of their values at z=0:!
Fc(0)! F+(0)! F-(0)! d Fc & Fc (d ) # & B11 0 0 # & Fc (0) #
= "(µ a + µ s )! Fc $ ! $ !$ !
z=0! dz Fc(0)! F+(0)! F-(0)!
z=0! $ F+ (d ) ! = $ B21 B22 B23 ! $ F+ (0) !
d F+ µ $ F (d ) ! $ B
= "(K + S )F+ + S ! F" + s (2 + 3 g )Fc % ' " % 31 B32 B33 !" $% F' (0) !"
Fc(z)! F+(z)! F-(z)! dz 4
d F" µs
= (K + S ) F" " S # F+ " (2 " 3g) Fc
Fc(z)! F+(z)! F-(z)! or! F(d ) = BF (0)
z!
dz 4 z! where Bpq(z) depend on µa, µs, and g, and
The total fluence rate is:! on specular reflection coefficients of the
z=d! collimated light rsp,c, the diffuse light rsp,d,
z=d!
! # ( z ) = Fc ( z ) + 2 " (F+ ( z ) + F! ( z ) ) as well as the internal diffusion reflection
coefficient rid !
See: van Gemert et al., "Optics of tissue in a multi-layer slab geometry", Lasers Life Sci 2(1), pp. 1-18,(1988)!
Multi-layers: Example "

Light distribution in multi-layers! Laser Illumination of Skin "
- A Four Layer Model!
Most biological tissues are stratified, i.e. consist of multi-layers
with different optical properties.! Estimation of the flux distribution in skin for a uniform and "broad"
Model: " collimated incident light Flux Fc0. !
For N multi-layers of thickness dj, characterised by µaj, µsj, Fc0= 1 W/cm2!
Source: "
gj (j=1,..., N) the flux vector FN is given by! Nd:YAG laser (1060 nm) ! Thickness d:!
N Fc0 = 1 W/m2! Epidermis! µa=10 cm-1! µs=23.33 cm-1! 0.0065 cm!
FN (d N ) = B N B N "1B N " 2 ...B 2 B1F1 (0) = ! B jF1 (0) = PF1 (0)
j =1
Upper Dermis! µa=3 cm-1! µs=47.67 cm-1! 0.03 cm!
Optical parameters of
where the matrix BN express the fluxes at z=d in terms of their Blood Plexus! µa=2 cm-1! µs=6 cm-1! 0.008 cm!
values at z=0:! skin at 1060 nm !
& Fc (d ) # & B11 0 0 # & Fc (0) # µa= absorption coefficient!
$ ! $ !$ ! Lower Dermis! µa=3 cm-1! µs=47.67 cm-1! 0.05 cm!
F (d ) = BF (0) or! $ F+ (d ) ! = $ B21 B22 B23 ! $ F+ (0) ! µs= scattering coefficient!
$ F (d ) ! $ B B32 B33 !" $% F' (0) !"
% ' " % 31
van Gemert et al., "Optics of tissue in a multi-layer slab geometry", Lasers Life Sci 2(1), pp. 1-18,(1988)!
Multi-layers: Example " Relation between KM coefficients and

Laser Illumination of Skin " the transport coefficients!
- A Four Layer Model!
It can be shown[1] that the relation between the transport coefficients
Fluxes for the skin model:! and the KM coefficients can be written as!
total fluence rate "
"=Fc(z)+2[F+(z)+F-(z)]!
collimated flux Fc(z)!
forward diffuse flux 2F+(z)! µa = " ! K µ s# = µ s (1 " g ) = $ ! S
+! backward diffuse flux 2F-(z)!
total diffuse flux 2[F+(z)+F-(z)]!
van Gemert et al., "Optics of tissue in a multi-layer slab [1] M.J.C. van Gemert and W.M. Star, "Relations between the Kubleka-Munk and the transport equation models for anisotropic
geometry", Lasers Life Sci 2(1), pp. 1-18 (1988)! scattering", Lasers Life Sci, 1(4), (1987), pp.287-298!
Relation between KM coefficients and Relation between KM coefficients and

the transport coefficients! the transport coefficients!
Relation between
microscopic transport
coefficients µa, µs, and g, " If scattering dominates absorption the KM coefficients can be
and the Kubelka-Munk directly related to the transport coefficients:!
coefficients K and S.!
The diagram shows:!
• curves of & and ' as a K ! 2µ a 3
S" µ s!
function of (1-W0).! 4
• approximations for & and '!
• values for K/S in relation to
µa
(1-W0) (horizontal axis)! (1 ! W0 ) = = the effective albedo!
µ a + µ s (1 ! g )
G. Wagnières, EPFL Thesis 1024 (1992) and M.J.C. van Gemert and W.M. Star, "Relations between the Kubleka-Munk and the transport equation models for
anisotropic scattering", Las Life Sci, 1(4), pp.287-298 (1987)!
3.4.3 Monte Carlo simulations to predict" Monte Carlo Approximation!

the light transport in turbid media!
• For situations where Diffusion theory breaks down,
the most useful method is Monte Carlo modeling!
" Computational technique which simulates multiple
scattering trajectories of individual photons through a
turbid medium!
" Each interaction is governed by the random processes
of absorption and scattering!
Monte Carlo simulations

Monte Carlo Simulations!
• No simplifications necessary!
• It is based on the transport equation.!
• Provides results for a specific case only -
• No further assumptions are made.! no analytical expressions!
• Microscopic parameters characterizing the • Photon statistics limit the signal-to-noise
medium are used (µa, µs, g, geometry).!
ratio!
• Extensive computer capacity is often
required!
Monte Carlo (MC) Simulations! Monte Carlo Simulations!

• Monte Carlo simulations are frequently used to model!
The outline of this presentation is:! !propagation of light in tissue.!
• This model is based on random walk, where a photon or!
• Introduction! !a photon package is traced through the tissue until it!
• Discuss how to randomize the parameters! !exits or is terminated due to absorption.!
• Briefly describe the calculations of step size • By repeating this process for a large number of photon!
and trajectory directions! !packages, it is possible to obtain statistics for these!
!physical quantities.!
• Many physical parameters of the photon package can!
!be logged, e.g. the distribution of absorption, exiting!
!position, time-of-flight etc.!
An example of random sampling! Monte Carlo Simulations!
Main advantage!
• no limitation concerning boundary conditions or!
spatial localization of inhomogeneities in the tissue!
Pond! Main disadvantage!

• problem of getting good statistics, particularly if the!
point of interest is located far away from the point of!
entry of the light and the scattering and absorption!
What is the area of the pond ?! coefficients are high.!
MC can handle any geometry! Monte Carlo Simulations!
• Monte Carlo simulations can be used for simulating

all kinds of particle transport. It has mainly been
developed for neutron transport in combination with
nuclear reactions.!
• For light transport in turbid media we consider
photons as neutral particles and neglect all wave
phenomena.!

It is as the name implies a method that relies on random
sampling of propagation variables from well defined
probability distributions!
Lets' "throw the dice” to determine:!

• The path length before a scattering or an absorption event!
occur!
• The scattering angle!
• ….!
Available Monte Carlo codes!

A public domain computer code (MCML), written by
Wang and Jacques treats the steady state case for a The outline of this presentation will be!
multi-layered medium.! • Introduction!
!http://omlc.ogi.edu/software/mc/index.html! • Discuss how to randomize the parameters!
!http://omlc.ogi.edu/news/dec98/mc321/index.html!
I) Selecting step size!
MCML can provide steady-state results for the II) Selecting deflection angle!
distribution of escaped as well as absorbed light in a III)Selecting azimuthal angle!
multi-layer geometry.!
Many developed versions based on this code are used

by various groups around the world.!
How to sample random variables 1! How to sample random variables 2!

Consider a probability density function (PDF) p(x) The probability that x will fall in the!
of x, which is such that:! interval [a, x1] is given by the
1!
cumulative distribution function
b (CDF) Fx(x1), defined as:!
! p ( x)dx = 1
a x1
Fx(x1 ) = " a
p(x)dx
With:! a ! x ! b
How to sample random variables 3! How to deduce x1 from )1!

"1 Fx(x1 )
By using a random number # 0
p(" )d" x1 CDF!
generator, one can obtain a random 1! = " a
p(x)dx
number ( in the range [0, 1].! = "1 1!
The probability density function for

this random number is 1 in the range ! !
[0, 1].! PDF!
The corresponding cumulative !
distribution function is:!
"1 This process is equivalent to equating the hatched areas under!
F" ("1 ) = ! p (" )d" = "1 0 ! "1 ! 1
the p()) and p(x) curves in the figure.!
The total areas under the p()) curve from 0 to 1 and under the!
0 p(x) curve from a to b are both equal to 1.!
How to sample random variables 4! Monte Carlo (MC) Simulations!

This means that the random number picked!
gives the integrated value of p(x), that is :!
The outline of this presentation will be!
x1 • Introduction!
"= # a
p(x)dx • Discuss how to randomise the parameters!
• Briefly describe the calculations of step
size and trajectory directions!
!
Light transport in tissue! Selecting the step size, s!
• The step size is in the interval [0, $].!
• The probability for an interaction in the interval [s1, s1+ds1]

• Path length! is µtds1. This means that the intensity I of the light that has
not interacted with the medium decreases in the interval ds1
• Scattering direction! by:!
• Absorption! dI(S1 )
= "µt I(S1 )
dS1
µt! =! µa!+! µs!
Selecting the step size, s! Selecting the step size, S!
The cumulative distribution function for free As we already know:!

path S is then (from Beer's law):!
S S
S1
P{S"S1} = 1# exp(#µt S1 ) = " p(S)dS "= # p(S)dS = # µt exp($µt S)dS =1$ exp($µt S)
0 0 0
yielding a probability density function! By rearranging this we get the final expression:!
! dP{S"S1} ! " ln(1 " ! ) " ln(! )

p(S1) = = µt exp(#µt S1 ) S= (=)!
dS
!1 µt µt
!
Phase function symmetry! Selecting the deflection angle!
• The phase function

• * is in the interval [0, +]!
has an axial • For convenience we will consider µ=cos(*)!
symmetry and depend • µ will be distributed in the interval [-1, 1]!
only on the angle *.! • For Mie scattering the Henyey- Greenstein is a good
• Often the phase approximation for the probability density function:!
function is expressed
as a function of the
cosine of the 1! g 2 1
scattering angle -p
p( µ ) =
2(1 + g 2 ! 2 gµ ) 3 / 2 # "1
p(µ)dµ = 1
(cos *)!
!
The Henyey-Greenstein function! Selecting the deflection angle!
By using equations derived earlier we get:!

µ µ
1# g 2
"= $ p(µ)dµ = $ 2 3/2
dµ
#1 #1 2(1+ g # 2gµ)
After solving this for µ we have:!

!
1 * 2
$ 1 " g2 '2-
µ= ,1 + g " & ) / for g%0!
2g ,+ %1 " g + 2g# ( /.
!
Selection of the propagation

Selecting the azimuthal angle!
parameters (Summary)!
The azimuthal angle is uniformly distributed!
within the interval [0, 2+]. Thus we have:!
$
1 $
"= % 2# d$ = 2#
0
This gives , in the form:! -$
! # = 2!"
Moving the photon package! Photon absorption!

• During each step some attenuation of the photon package
• Specify the step size S as described! weight occur due to absorption.!
• For a photon located at (x, y, z), traveling a distance S in the µa
direction (µx, µy, µz), the new coordinates (x', y', z') are given !Q = W
The deposited energy is! µt
by:!
x'= x + µx S
y'= y + µy S !W = W
µS
and the new photon weight is!
µt
z'= z + µz S
(note that &Q + &W = W)!
!
Launching a photon package on a Internal reflectance or escape!

mismatched boundary!
• For each step one has to check if the photon
package crosses a boundary (internal/external).!
• We start with launching a photon package • If this is the case one has to check for internal
assigned with a weight W, equal to unity.! reflectance and/or escape. This is calculated using
the Snell and Fresnel laws.!
• Some specular reflections will occur if there is a
• The escaped fraction adds to the result file and
mismatched boundary: W=1-R!
reduces the weight of the photon package left.!
Photon package termination! Photon package termination!

A technique called Roulette is used to terminate the photon when!
W/m = W' < Wthreshold!
• After a certain number of scattering events the!
A random number, rnd, in the interval [0,1] is generated and the
!remaining photon package weight will be low.! photon is updated as follows: !
• It is certainly a waist of time to follow this!
If rnd " 1/m then, mW # W
!package once it has a weight lower than a preset!
!threshold Wth.! If rnd > 1/m then, W = 0
• The algorithm used to have an efficient MC! This gives the photons with a weight of less than the threshold a
chance of 1 out of m to survive (e.g. m=10). The goal is to conserve
!code, is to use a technique called roulette.! energy and to terminate in an unbiased manner.!
!
MC models enables simulate the generation Important applications!

of fluorescence in layered tissue structures!
• Dosimetry in connection to PDT of!

!superficial or deep seated lesions!
• Fluorescence diagnostics of early!

!malignant tumours!
• etc!
3. Tissue Optics! 3.4 Approximations of the RTE!
Example!
The propagation of a 1-mJ laser impulse into
tissue (M.-C. simulation, µa = 0.1 cm-1, µs =
100 cm-1, g = 0.9, n = 1.37). In each figure, a
snapshot of the light distribution is depicted
as contour curves which indicate iso-density
lines in units of J/cm3. (a) At 0.46 ps, the
leading edge of the impulse has traveled 100
µm. The dot at r = o and z = 100 µm
represents a delta-function that describes the
unscattered energy density at 0.46 ps. (b) At
2.3 ps, the leading edge is at 500 µm. (c) At
46 ps, the leading edge is at 1 cm, but it is so
low in value that it is not shown!
Georges Wagnières, ISIC, EPFL! Photomedicine, spring 2012! 79!

3 4 Approxim of The RTE 79

Uploaded by

Copyright:

Available Formats

3 4 Approxim of The RTE 79

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

3 4 Approxim of The RTE 79

Uploaded by

Copyright:

Available Formats

3. Tissue Optics! 3.4 Approximations of the RTE! 3. Tissue Optics! 3.4 Approximations of the RTE!

Radiative Transport Equation! Diffusion Approximation!

• Can be solved analytically for special cases or more

Gains: scattering + source • Validity of diffusion equation is limited to tissue cases

Diffusion approximation! Radiant Energy Fluence Rate " (r,t)!

The diffusion approximation has been used extensively in

Diffusion approximation! Solutions of the Diffusion Approximation"

! !where J is the photon current!

• Interstitial illumination with a cylindrical light distributor!

Interstitial "point" source in an infinite

This Green’s function can be used to solve problems

Steady State Diffusion! Solution to stady state diffusion!

This yields a solution for an infinite homogenous

Point source solution" The Time-independent Fluence Rate!

n1 #$ Incident light of power: P

Collimated “Broad” illumination Collimated “broad” illumination

Irradiance: "0 Diffusion

Interstitial illumination with an infinitely long"

K0 and K1 are modified Bessel functions of the second kind!

3.4.2 The Kubelka-Munk Theory! Radiance!

Kubelka-Munk Theory! The Kubelka-Munk Theory!

Measurement of the Reflection and Transmission! KM theory for collimated flux!

Three flux theory!

Three flux Approach! The three flux approach!

Multi-layers: Example "

Multi-layers: Example " Relation between KM coefficients and

Relation between KM coefficients and Relation between KM coefficients and

3.4.3 Monte Carlo simulations to predict" Monte Carlo Approximation!

Monte Carlo simulations

Monte Carlo (MC) Simulations! Monte Carlo Simulations!

An example of random sampling! Monte Carlo Simulations!

Pond! Main disadvantage!

MC can handle any geometry! Monte Carlo Simulations!

• Monte Carlo simulations can be used for simulating

Monte Carlo Simulations!

Lets' "throw the dice” to determine:!

Available Monte Carlo codes!

Many developed versions based on this code are used

How to sample random variables 1! How to sample random variables 2!

How to sample random variables 3! How to deduce x1 from )1!

The probability density function for

How to sample random variables 4! Monte Carlo (MC) Simulations!

Light transport in tissue! Selecting the step size, s!

• The step size is in the interval [0, $].!

• The probability for an interaction in the interval [s1, s1+ds1]

Selecting the step size, s! Selecting the step size, S!

The cumulative distribution function for free As we already know:!

! dP{S"S1} ! " ln(1 " ! ) " ln(! )

Phase function symmetry! Selecting the deflection angle!

• The phase function

The Henyey-Greenstein function! Selecting the deflection angle!

By using equations derived earlier we get:!

After solving this for µ we have:!

Selection of the propagation

This gives , in the form:! -$

Moving the photon package! Photon absorption!

Launching a photon package on a Internal reflectance or escape!

Photon package termination! Photon package termination!

MC models enables simulate the generation Important applications!