BMED322

Functional Biomechanics
Lecture 4
Anthropometry
The study of humans
Anthropology
>
-
 Anthropometry
• Anthropometry
• ‘The measurement of humans’
• Measurement of the physical characteristics of segments of
the human body
• Dimensions motion
rotation

• Inertial properties (Mass, Moments of Inertia)

• Frequently used simplifying assumptions:

• Body segments behave as rigid objects
• Mass distribution is similar among members of a particular
population

2
 History of Anthropometry
The Ancient Roman architect Vitruvius (1st Century BC), wrote that in the human body:

A palm is the width of 4 fingers

A foot is the width of 4 palms
A cubit is the width of 6 palms
A man's height is 4 cubits (and thus 24 palms)
The length of a man's outspread arms is equal to his height
The distance from the hairline to the bottom of the chin is 1/10 of a man's height
The maximum width of the shoulders is 1/4 of a man's height
The distance from the elbow to the tip of the hand is 1/5 of a man's height
The distance from the elbow to the armpit is 1/8 of a man's height
The length of the hand is 1/10 of a man's height
The distance from the bottom of the chin to the nose is 1/3 of the head length
The distance from the hairline to the eyebrows is 1/3 of the length of the face
The length of the ear is 1/3 of the length of the face

3
 Anthropometry in Human Movement
• Divide the body into multiple
segments

• Each segment has:

• Segment Length
• Segment circumference
• Mass and Centre of Mass
• Moment of Inertia
• Joint Centers of Rotation
• Muscle Origin and Insertion Points
• Muscle Length and Cross-Section.

4
 Quantifying Segment Parameters
Several methods may be used including
• Cadaver studies Dead body
>
-

• Direct kinematic measurements

• Imaging techniques

5
 Direct Kinematic Measurements
Locating the centre of gravity using a balance board
-

O
weight of
the board
𝑥𝐶𝑀
Red
𝑥𝐶𝑀 = Wbxp-R .
d = 0 Wan + Wacm =

Ri
=

6
 M
I Ma =
=

Ld)

Direct Kinematic Measurements

Moment of inertia measurement using a torsional pendulum

Restoring torque, 𝜏 = −𝜅𝜃 = 𝐼𝛼

𝑇2𝜅
𝐼= 2
4𝜋

7
 Imaging Techniques
Various radiation techniques are used to scan the
living body
• Gamma mass scanning
• Dual-energy X-ray absorptiometry (DEXA)
• Magnetic resonance imaging (MRI)
• Photogrammetry

8
 Mathematical Modelling
• Pioneered by Hanavan, who assumed:
• Uniformly distributed mass
• Rigid body segments
• Each segment described by a geometrical
shape
• 3 segments/limb + 2 for the trunk + head segment
• Segments modelled as frustra (slices) of
right-circular cones
• Hands as spheres
• Head as ellipsoid

• Preferably measure dimensions directly from each

individual for avoiding bias (gender/racial)

9
Segment Measurements: Drillis and Contini

Body segment lengths are expressed as a fraction of the body height, H

10
 Whole-Body Density
• Mass per unit volume (kg/m3)
• Human body consists of many types of tissue, each
with a different density.
• Average density is a function of body build called
somatotype
• Endomorph – short and fat
• Mesomorph – stout and muscular
• Ectomorph – tall and thin
• Simplest way is to measure weight and volume of the
body (Archimedes principle) – not convenient (or dry!)

11
 Whole-Body Density
• In the 1960s, Contini developed an expression for
body density d as a function of c ponderal index
(similar to the commonly-used BMI)

ℎ
Ponderal index 𝑐 = 3 (h=height and m = mass)
𝑚

Density = 0.69 + 0.9𝑐 (kg/L)

12
 Whole-Body Density
Example: Calculate the whole-body density of an adult
with h = 1.78 m and m = 77.3 kg

Ponderal index c = h/m1/3 (h=height and m = mass)

Density = 0.69 + 0.9 c (kg/L)

C
= y
=
0
.

4 M/Ke

4) 1 4149/L
Density 69 0 9(0 =
.

0
.
.

=
.
+

13
 Segment Densities
• Each body segment has a unique • Can express the mass of each
combination of bone, muscle fat and segment as a percentage of the
other tissue and the density within a total body mass because the
given segment is not uniform. mass of individual segments
increases with total body mass.
• In general, limbs are denser than the
• Individual segment densities
trunk (due to higher ratio of bone). increase as whole-body density
increases. Distal segments have
higher densities than proximal
segments.

14
Automated Landmark Detection

Bazarevsky et al., “BlazePose: On-device Real-time Body Pose tracking,” 2020

15
Automated Landmark Detection

Bazarevsky et al., “BlazePose: On-device Real-time Body Pose tracking,” 2020

16
Automated Landmark Detection

Bazarevsky et al., “BlazePose: On-device Real-time Body Pose tracking,” 2020

17
Automated Landmark Detection

Bazarevsky et al., “BlazePose: On-device Real-time Body Pose tracking,” 2020

18
Automated Landmark Detection

Bazarevsky et al., “BlazePose: On-device Real-time Body Pose tracking,” 2020

19
Machine Learning & Neural Networks
There are various flavors of Machine Learning (ML)
methods:
• Supervised ML: Training model using labelled data
• Unsupervised ML: ‘Finding patterns’ in unlabelled data

20
Machine Learning & Neural Networks
Machine Learning (ML) tasks include
Classification
see

Fruit Sorting

Regression
use

Density = 0.69 + 0.9 c (kg/L)

21
Machine Learning & Neural Networks
• Nonlinear regression/classification problems

22
Machine Learning & Neural Networks

Output

23
Machine Learning & Neural Networks
Model training involves ‘learning’ model weights

Once trained, the weights are ‘fixed’ and the model

is run in ‘inference’ mode

Various Python packages exist such as PyTorch and

Tensorflow for convenient model training

24