Measuring Biological Matter_2012
-
Upload
christian-lovato -
Category
Documents
-
view
214 -
download
0
Transcript of Measuring Biological Matter_2012
-
7/31/2019 Measuring Biological Matter_2012
1/6
What is biological matter? It is group of atoms (matter) that has (or had) the property of being alive.
We don't go in the definition of wich mean "being alive", but is a common notion, learned by examples.
Some examples are
A tree
A cat
A cell
A tissue
Christian Lovato
Measuring biological mattersabato 24 aprile 2010
13.42
Info-Med.one (On 19-01-2011) Page 1
-
7/31/2019 Measuring Biological Matter_2012
2/6
Measure:
The quantity,size,weight,distance or capacity of a substance compared to a designated standard.
Incollato da
we can say "give a number (or a set of numbers) indicating something"
Information theoryInformation theory recognizes that all
data are inexact and statistical in
nature. Thus the definition of
measurement is: "A set of observationsthat reduce uncertainty where the
result is expressed as a quantity."[3].
This definition is implied in what
scientists actually do when they
measure something and report both the
mean and statistics of the
measurements. In practical terms, one
begins with an initial guess as to the
value of a quantity, and then, using
various methods and instruments,
reduces the uncertainty in the value.
Note that in this view, unlike the
positivist representational theory, all
measurements are uncertain, so instead
of assigning one value, a range of values
is assigned to a measurement. This also
implies that there is a continuum
between estimation and measurement.
Incollato da
Measure of an artifact
With some instruments
We can define a diameter, an height, without ambiguities.
After, we can answer some questions like:
how many mugs can enter in this box?
A mug
We obtain a set of
Informative (useful)
measures
A human
Info-Med.one (On 19-01-2011) Page 2
http://en.wiktionary.org/wiki/quantityhttp://en.wiktionary.org/wiki/quantityhttp://en.wiktionary.org/wiki/sizehttp://en.wiktionary.org/wiki/sizehttp://en.wiktionary.org/wiki/weighthttp://en.wiktionary.org/wiki/weighthttp://en.wiktionary.org/wiki/distancehttp://en.wiktionary.org/wiki/capacityhttp://en.wiktionary.org/wiki/substancehttp://en.wiktionary.org/wiki/standardhttp://en.wiktionary.org/wiki/standardhttp://en.wiktionary.org/wiki/measurehttp://en.wiktionary.org/wiki/measurehttp://en.wikipedia.org/wiki/Information_theoryhttp://en.wikipedia.org/wiki/Measurement#cite_note-2http://en.wikipedia.org/wiki/Measurement#cite_note-2http://en.wikipedia.org/wiki/Meanhttp://en.wikipedia.org/wiki/Statisticshttp://en.wikipedia.org/wiki/Positivisthttp://en.wikipedia.org/wiki/Estimationhttp://en.wikipedia.org/wiki/Measurementhttp://en.wikipedia.org/wiki/Measurementhttp://en.wikipedia.org/wiki/Measurementhttp://en.wikipedia.org/wiki/Estimationhttp://en.wikipedia.org/wiki/Positivisthttp://en.wikipedia.org/wiki/Statisticshttp://en.wikipedia.org/wiki/Meanhttp://en.wikipedia.org/wiki/Measurement#cite_note-2http://en.wikipedia.org/wiki/Information_theoryhttp://en.wiktionary.org/wiki/measurehttp://en.wiktionary.org/wiki/standardhttp://en.wiktionary.org/wiki/substancehttp://en.wiktionary.org/wiki/capacityhttp://en.wiktionary.org/wiki/distancehttp://en.wiktionary.org/wiki/weighthttp://en.wiktionary.org/wiki/sizehttp://en.wiktionary.org/wiki/quantity -
7/31/2019 Measuring Biological Matter_2012
3/6
in the biological domain the situation is usually different.
We can measure as we do with artifacts, but we have a lot of problems, for example in defining what
and how to measure. In most cases biological systems cannot "fit" into a determined number of
measures.
For example one of the common measures of the body is stature. Measuring height is simple. A little bitmore complicated than measuring the height of a mug but we can obtain a number that gives the
stature of a person. However there are some problems once we want to use this number.
For instance, there is a problem with this number's precision. We know that the stature of a person
changes during life, so that the temporal validity of this number is limited. And if we want to be more
precise, the height of a person decrease of about 1% during the daytime and increase during the sleep.
[http://www.informaworld.com/smpp/content~db=all~content=a785169438 ]
Another order of problems that emerges is about the definition of the measure, i.e. how to deal with
hairs during the measurement: if we compress them we obtain a certain number, if we don't we have
another, and if we cut them and measure on the scalp we have yet another measurement.
None of them is the "real" height, every measure has a different biological meaning, if not a clinic one.
All we have said about stature is valid for almost all the measures we can think to perform.
These problems derive, as last resort, from the fact that biological systems (and thereby shapes) are
complex, and we don't have a well defined geometrical, or mathematical, model of what a "human
body" is.
While we can have few ambiguities in measure and use the measures of the mug, we really have a lot of
ambiguities in measuring, and in using the measures of biological systems, because of their complexity.
How to approach this complexity? Or, more simply, how to get objective measurements of biological
systems?
Firstly, we need a model of the human body. Every measure is
related to a model that express some of the relationships of the
subject.
While for the measures of a mug we can simply define one (the
measure of "diameter" is strongly connected with the notion ofthe "cylinder" that forms the geometrical model of the mug) in
the biological case, this model is more complex and should be
explicitly defined, for every set of measures we want to obtain.
Secondly, we can manage the complexity of the "human body", and of the model that should be
derived from it, by means of successive refinements. We can start from simple general models, and
"deepen" them into more complex and precise models that use the information obtained from the
simpler ones, in a process can be compared to the successive approximation of an optimization
algorithm.
These considerations lead to a general framework in wich a such model can be constructed, and in wich
The model is usually in the mind of
the operator, and must be shared
with the agent that uses the
measurement.
Different people have usually similar
(not equal) models of the same
subject.
Info-Med.one (On 19-01-2011) Page 3
http://www.informaworld.com/smpp/content~db=all~content=a785169438http://www.informaworld.com/smpp/content~db=all~content=a785169438http://www.informaworld.com/smpp/content~db=all~content=a785169438 -
7/31/2019 Measuring Biological Matter_2012
4/6
the notion of successive approximation is implemented.
We can expand this framework by including new data sources. We can also decide to set up interactive
measures for those blocks for which automatic measures are not readily avaiable, or the implemented
algorithms fail for some specific data.
Other interesting features and properties of the method can be:
Substitution of algorithms when better (faster, or more robust) algorithms become avaiable, alterating
only a block, with eventual minor changes in neighbours blocks
Measures on request: if an automatic pipe is heavy to process as a whole, only the needed blocks can be
activated (backpropagation of request)
Associate to every measure some indicative label (from medical domain )
Reusability of well-behaving algorithms within the system at different levels/scales, where previous
stages provide the context to deeper ones. This frame fits naturally in a multirisolution approach, i.e.
different levels could use data at different resolutions.
In this very general frame we take row data as input, preprocess them in order to achieve a good
precision against acquisition errors, and process them at different stages, or levels.
At every stage we refine the explicit information about the particular examined dataset .
The key is that we can use information extracted at early levels to simplify (make more efficient, faster,
simpler to implement) the tasks of the deeper levels.
Info-Med.one (On 19-01-2011) Page 4
-
7/31/2019 Measuring Biological Matter_2012
5/6
In general, strenght of this approach is the good coupling it can give between "medical knowledge" and
the implemented model, because the measures are guided from the medical needings, and the
representation can give useful medical informations at all levels of frame, from the simplest ones (that
give general information) to the deeper levels, that extract more specific or contextualized informations.
This feature is necessary in order to obtain immediatly useful and informative measures of the subject.
What's already done at level 1
Example (our anthropometric application)
Info-Med.one (On 19-01-2011) Page 5
-
7/31/2019 Measuring Biological Matter_2012
6/6
Improvements:
allows analysis of fine structures; i.e. hands, that in the original mesh has motion artifacts
Advanced hole closing -> refinement in volume measurement
Better mesh preprocessing ->
Add algorithms for interesting points on mesh surface, in order to make segmentation more robust, or
make it work in presence of topological variations due to different subject poses
Planning to improve the stick fitting algorithm.