Saturday,
4th August 2012
Report:
·
Aiming
to analyze the audio output available from Ultrasound machines during
Continuous Wave or Pulse Wave Doppler Examination for detection of placental
insufficiency.
ü Instrumentally possible to get
audio output from an ultrasound machine.
·
Need
to perform a Fast Fourier Transform
Analysis on the audio signal to fetch meaningful information from it, that
is to remove the background noise..
?
Has
that been done before?
?
Where
can the Ultrasound CW/PW Doppler Audio output Database be found? If no database
available, is the project feasible?
?
What
is the audible output actually? Is it the same frequency signal that is sent
through the ultrasound transducer? Or is it a modulated signal of the
ultrasound converted to audio frequency?
·
ultrasound
frequency range: greater than 20KHz and audio frequency(AF) or Acoustic range:
20Hz to 20KHz.
·
Medical
use : 1 – 6 MHz for Deeper organs and 7 – 18 MHz for superficial organs.
2
– 7 Mhz for adults and upto 10/12 for younger patients
!
Need
more published papers to confirm feasibility.
Sunday,
5th August 2012
New
established facts:
·
audio
output are the sounds generated by pitch changes of the ultrasound waves
·
The
changing velocities (frequencies) are converted into audible sounds and, after some
processing, are emitted from speakers placed within the machine.
·
High
pitched sounds result from large Doppler shifts and indicate the presence of
high velocities, while low pitched sounds result from lesser Doppler shifts.
Flow direction information (relative to
the transducer) is provided by a
stereophonic audio output in which flow toward the transducer
comes out of one speaker and flow
away from the transducer.
·
The
audio output also allows the operator to easily differentiate laminar from
turbulent flow.
Laminar flow produces a smooth,
pleasant tone because of the uniform velocities. Turbulent flow,
because of the presence of many
different velocities, results in a commonly high-pitched and
whistling or harsh and raspy
sound.
·
The
trained ear can readily appreciate minor changes in spectral composition more
readily than the eye, given the same information displayed graphically.
·
The
major limitation of audio Doppler outputs is the requirements for subjective
interpretation and the lack of a permanent objective record.
·
The
audio output from a Doppler machine is not the same as that received by a
stethoscope or a phonocardiogram. The sounds detected with a stethoscope are
transmitted vibrations or pressure waves from the heart and great vessels that
are believed to be the result of rapid accelerations and decelerations of
blood.
·
The
Doppler audio output, in contrast, is an audible display of the Doppler
frequency shift spectrum produced by red cells moving in the path of the
ultrasound beam. It is a sound produced by the Doppler machine that does not
occur in nature and, therefore, it does not originate in the heart.
?
The
deciphering, analyzing and presenting vast amounts of returning data into
various velocity components is popularly done using either of the 2 methods.
Which one to go for? Why?
1. FFT
2. Chirp-Z
!
Need
documentation on the instrumentation part of how the audio signal is produced
from ultrasound returning waves.
!
More
papers needed which relate the audio and imaging outputs.
?
How
much more information or how much less information does the audio provide
compared to ultrasound image output?
?
Which
Doppler method to use? Continuous wave or Pulsed Wave? Even though both have
their advantages and disadvantages in terms of imaging, what are pros and cons
in terms of audio signal analysis?
?
What
are the other audio controls available on the ultrasound Doppler machine in
order to “tune” to an optimum audio output other than ultrasound controls?
Monday,
6th August 2012
·
Need
to established the connectivity in the project development process.
Back-end à Activity à Front end
·
Back
end consists of the signal acquisition problem:
Instrumental Problem:
?
Is
a inbuilt recording device available in the ultrasound machines?
?
Else,
Are Doppler audio signals database available anywhere?
?
Else,
can we record the output?
Using the recorded signal:
?
What
format is the signal available?
?
Which
software do we use to analyze it? Matlab?
·
Activity
consists of meaningful conversion problem
?
Which
conversion technique to use? Advt and disadvt of each technique.
?
Which
all cases to study?
?
Which
trimester to concentrate on for accurate detection?
·
Front
end consists of the result and interpretation
Tuesday,
7th August 2012
·
Met
up with Gynac, Dr. Govind Kamat who referred me to Dr. Sanjay Sardesai from
Ultrasound Clinic
·
Doubts
Cleared :
Þ
Audio
signals generated are stored on the hard disk, which are accessible.
Þ
Audio
signal gives the same information as the imaging signals. Hence the project is
feasible.
!
Imaging shows the exact internal
organ location where transducer is operating on, the information unavailable in
audio. ç
!
Database
bank availability still unclear
·
Rest
questions remain the same.
Saturday,
11th August 2012
·
Submitted
a very ununderstandable project synopsis yesterday. Here’s the jist:
“Problem
Statement:
The Doppler audio output is an audible
display of the Doppler frequency shift spectrum produced by red cells moving in
the path of the ultrasound beam. The trained ear can readily appreciate minor
changes in spectral composition more readily than the eye, given the same
information displayed graphically. The major limitation of audio Doppler
outputs is the requirements for subjective interpretation and the lack of a
permanent objective record.
Motivation:
If audio Doppler output could be
analyzed and characterized for a particular condition, to generate statistics
that could reliably correlate the audio and image output of the ultrasound
machine, the diagnosis procedure can be enhanced or simplified.
Brief Background:
The ear and the brain are extraordinarily
well adapted to selecting and interpreting complex sound patterns immersed in a
background of noise and irrelevant data. But for this, the listener needs to
learn to recognize clinically important sounds in Doppler signals. Thus, the
interpretation of the audio signal is subjective in nature. The audio Doppler
output obtained for a particular medical condition could be analyzed using a
suitable adaptive signal processing technique for feature extraction, and its
interpretation can be correlated with image Doppler output for reliability. The
result then could be used as a diagnostic tool for audio Doppler output of that
particular medical condition, reducing the subjective nature of its
interpretation.”
·
Need
to divide the project into stages of achievements, for this, need to understand
the procedure I’d be adopting
·
Official
duration allotted: 3rd July 2012 to 30th May 2013.
·
Total
11 months, more than 1 month already gone, 10 months remaining.
·
Considering
2 to 3 achievements every month, dividing the project into 8 check points.[CP 1 - 8]
·
As
discussed before, project is divided into 3 stages:
Þ
Study, Signal Acquisition and
Feeding into system:-
For this, need to figure out the below(Well begun is half done): Duration: 2.5
months
§
Theory
behind Clinical Ultrasound Doppler Audio Signal Generation, its interpretation
in diagnosis
§
Ultrasound
Doppler instrumentation and its working
§
What
are the procedure to be followed in order to extract information from the
signal? CP 1 –
End of August
§
Availability
of the signal: Where, what format?
§
Which
software will be appropriate for this signal processing?
§
Getting
the licensed copy of the software and acquire compatible format audio data CP 2 – End of September
§
Feed
to the system. Stage Complete? Verified? CP 3 – 15th of october
Þ
Signal Processing for Feature
Extraction(Heart
of the project): Duration: 7 months
§
Study
of various techniques to be used CP 4 – 15th of November
§
Implement
CP 5 – End of
November
§
Trial
§
Observation
§
matching
with ultrasound imaging for reliability
§
repeat
for satisfactory number of trials in various conditions CP 6 – 15th of
May
Þ
Interpretation of the acquired
result(All’s
well that ends well): Duration: 0.5 months
§
Interpretation
of result
§
Application
§
Future
Scope CP 7 –
End of 3rd week of May
§
Presentation
§
Documentation
CP 8 – End of
May
Sunday,
12th August 2012
·
Yesterday,
read through the first part of the Ultrasound book Xeroxes that Dr. Sardesai
insisted I refer to, called, Clinical Applicatioons of Doppler Ultrasound by
Kenneth J W Taylor. Tumbled upon the fact that the same signal that is heard is
shown on the ultrasound machine as a waveform. If it is already shown as a
waveform, then, what am I trying to do using the audio output? I am intending
to get the information given by the audio signal and convert it into visual
aid, which is actually the same as that already available visually. PROBLEM?
·
NO. Coz my intention is, to the
program software detect the changes in audio signal in the same way that a
doctor does and give the result, WITHOUT
THE AVAILABILITY OF A VISUAL AID. So, only a small ultrasound Doppler audio
machine will be sufficient to detect a problem, and wont cost 600 rs per
ultrasound.
·
I
would currently be working on detection of a particular condition, (which
condition??? depends on whether transducer location can be confirmed) and the
list of conditions can then be increased upon by further research.
·
Handheld
audio Doppler instruments are already existing. They must be used with some
intention. Do they help in detection of any abnormality(That’s a better word)?
other than giving the audio output? Will have to check the working and
intentional usage of a handheld Doppler device.
·
My
objective is to be able to reliably detect an abnormality from audio signal
using digital signal processing techniques.
·
Deducing
that my project basically deals with the audio output of handheld Fetal Doppler
machine, the problem that arises is the acquisition of the audio signal, since
not sure if handheld machine has a storage facility. That poses the same
project feasibility problem as before.
STUCK‼
